Last-mile vehicle: Digital and physical prototyping of high-performance cargo bikes PDF Free Download
1 / 124
/124
100%
ARENBERG DOCTORAL SCHOOL
FACULTY OF ENGINEERING TECHNOLOGY
Last-mile vehicle
Digital and physical prototyping of
high-performance cargo bikes
Jordi D’hondt
Dissertation presented in partial
fulfilment of the requirements for the
degree of Doctor of Engineering Technology
(PhD): Mechanical Engineering
May 2024
Supervisors:
Prof. Dr. Ir. Peter Slaets, promotor
Prof. Ir. Marc Juwet, co-promotor
Prof. Dr. Ir. Eric Demeester, co-promotor
Last-mile vehicle:
Digital and physical prototyping of
high-performance cargo bikes
Jordi D’hondt
Dissertation presented in partial fulfilment of the requirements for the degree
of Doctor of Engineering Technology (PhD)
May 2024
Examination committee:
Prof. Dr. Ir. Peter Slaets (supervisor)
Prof. Dr. Ir. Eric Demeester (co- supervisor)
Prof. Ir. Marc Juwet (co- supervisor)
Prof. Dr. Ir. Jan Cappelle (assessor)
Dr. Ir. Sander Vandenberghe (assessor)
Prof. Dr. Ir. Johan Ceusters (chair)
Prof. Dr. Ir. Nobby Stevens (secretary)
Prof. Dr. Ir. Li Zhou
© 2024 KU Leuven – Faculty of Engineering Technology
Uitgegeven in eigen beheer, Jordi D’hondt, Gebroeders De Smetstraat 1, Gent, België
Alle rechten voorbehouden. Niets uit deze uitgave mag worden vermenigvuldigd en/of openbaar
gemaakt worden door middel van druk, fotokopie, microfilm, elektronisch of op welke andere
wijze ook zonder voorafgaande schriftelijke toestemming van de uitgever.
All rights reserved. No part of the publication may be reproduced in any form by print, photoprint,
microfilm, electronic or any other means without written permission from the publisher.
i
Acknowledgement
I’m proud to say: I did it! However I couldn’t do it without the support of my
surroundings. Therefore I would like to take this moment to thank some of
you in particular.
First off all, I would like to thank my love Ellen for the support during my
research and for giving me the time to finish my dissertation. Without you this
wouldn’t be possible. I also want to thank Cyriel for being calm while sitting
on my lap during your first months while I was finishing my dissertation.
Marc Juwet for giving me the opportunity to pursue my PhD. It’s been a
challenging and inspirational journey. Thank you for your guidance and the
support.
My colleagues who supported me during the journey. There was always an
enjoyable working environment and I could always count on your support.
With a special thanks to the other researchers: Jannes, Pieter, Ward, Theo and
Robin. You guys kept me motivated.
To the esteemed members of the research committee, Thank you for sharing
your knowledge and giving feedback when needed. Your expertise, insights
and constructive feedback have been invaluable throughout this process and
aided me to strive for perfection.
To my cherished friends and family. Completing this dissertation wasn’t
possible without your support.
And last but not least, a big shootout to everyone that I forgot to mention.
Every gesture big or small has been highly appreciated.
ii
iii
Summary
The Courier, Express and Parcel service sector has experienced consistent
growth, primarily attributed to the increasing popularity of E-commerce. In
Belgium alone, the year 2020 saw the delivery of 336 million parcels. The
last-mile delivery, which represents only a small portion of the travelled
distance but approximately half of the logistical chain cost, presents
significant challenges. Delivery of on average 0.1 parcels per person daily is
primarily done by vans. However, this reliance contributes to pollution and
congestion issues in city centres, prompting municipalities to implement
measures such as circulation plans, low-emission zones, and car-free zones.
To address these challenges, this doctoral dissertation proposes the utilization
of cargo bikes as a potential optimization solution for last-mile delivery. Cargo
bikes offer advantages such as reduced congestion, the ability to utilize bicycle
lanes, easier parking, and a lower ownership costs. Moreover, governments
are actively promoting environmentally friendly alternatives. Despite these
benefits, cargo bikes face limitations in cargo capacity. To overcome this
obstacle, the research introduces a novel cargo bike design, specifically
engineered to enhance cargo space and handling qualities. Beginning with a
comprehensive market research phase to identify vehicle requirements and
assess existing cargo bike models, the development process progresses to
prototype creation.
The resulting prototype is a three-wheeled vehicle featuring a conventional
cyclist module and a cargo area capable of transporting a euro pallet up to 200
kg between the two rear wheels. The design allows for independent tilting of
the driver, facilitating manoeuvrability. Propulsion is achieved through an
electrical drivetrain system, comprising a pedal generator and two hub motors.
A sophisticated control loop is developed to simulate natural pedalling
dynamics and propel the vehicle efficiently. This is crucial as conventional
pedalling power varies according to pedal orientation. The prototype also
undergoes rigorous testing to evaluate its handling qualities, including
predetermined manoeuvres to assess its performance.
Ultimately, this research culminates in the development of a novel cargo bike
with increased cargo capacity and improved handling, offering a promising
solution to the challenges of last-mile delivery in urban environments.
iv
Samenvatting
De sector voor koerierdiensten heeft een consistente groei doorgemaakt. De
toenemende populariteit van E-commerce heeft hierbij geholpen. Alleen al in
België werden in 2020 336 miljoen pakketten bezorgd. De last-mile delivery
vertegenwoordigt slechts een klein deel van de totaal afgelegde weg, maar
kost ongeveer de helft van de logistieke keten. Dit brengt aanzienlijke
uitdagingen met zich mee. Het leveren van 0,1 pakket per persoon per dag
wordt voornamelijk uitgevoerd met bestelwagens. Dit draagt bij aan
vervuilings- en fileproblemen in stadscentra, waardoor stadsbesturen
maatregelen nemen zoals circulatieplannen, lage-emissiezones en autovrije
zones.
Om deze uitdagingen aan te pakken, stelt dit doctoraat voor om cargo fietsen
te gebruiken als een mogelijke optimalisatie voor bezorging in stadscentra.
Cargo fietsen bieden voordelen zoals verminderde file, het vermogen om
fietspaden te gebruiken, gemakkelijker parkeren en lagere kosten. Bovendien
bevorderen overheden actief milieuvriendelijke alternatieven. Ondanks deze
voordelen hebben cargo fietsen beperkingen met betrekking tot
laadvermogen. Om dit obstakel te overwinnen, introduceert het onderzoek een
nieuw ontwerp van een cargo fiets die specifiek ontwikkeld is om de
laadruimte en de hanteerbaarheid te verbeteren. Beginnend met een
uitgebreide marktonderzoeksfase om voertuigvereisten te identificeren en
bestaande cargo fietsmodellen te beoordelen, vordert het ontwikkelingsproces
naar de creatie van een prototype.
Het prototype is een driewielig voertuig met een conventioneel
bestuurdersdeel en een laadruimte geschikt voor het vervoeren van een 200 kg
geladen euro-pallet tussen de twee achterwielen. Het ontwerp maakt
onafhankelijk kantelen van de bestuurder mogelijk, waardoor de
manoeuvreerbaarheid wordt vergemakkelijkt. Voortstuwing wordt bereikt
door middel van een elektrisch aandrijfsysteem, bestaande uit een
pedaalgenerator en twee naafmotoren. Een geavanceerde controle lus is
ontwikkeld om een natuurlijke pedaaldynamiek te simuleren en het voertuig
efficiënt aan te drijven. Dit is cruciaal omdat het conventionele trappen
varieert afhankelijk van de pedaaloriëntatie. Het prototype ondergaat
rigoureuze tests om de hanteerbaarheid te evalueren, inclusief vooraf bepaalde
manoeuvres om de prestaties te beoordelen.
Uiteindelijk resulteert dit onderzoek in de ontwikkeling van een nieuwe cargo
fiets met een verhoogd laadvermogen en verbeterde hanteerbaarheid. Dit biedt
een veelbelovende oplossing voor de uitdagingen van last-mile bezorging in
stedelijke omgevingen.
v
List of abbreviations
ABS Anti-lock Braking System
B2B Business to Business
B2C Business to Customer
BLDC Brushless Direct Current
C2C Customer to Customer
CAGR Compound Annual Growth Rate
CEP Courier, Express and Parcel
CVT Continuous Variable Transmission
e-bikes electric bicycles
EPAC Electric Pedal Assisted Cycle
ESC Electric Stability Control
IAV Instantaneous Angular Velocity
LMV Last-Mile Vehicle
MFDD Mean Fully Developed Deceleration
PSHD Pedal Series Hybrid Drivetrain
QFD Quality Function Deployment
SOC State Of Charge
SPM Smart Personal Mobility
vi
vii
Table of Contents
Acknowledgement ........................................................................................... i
Summary........................................................................................................ iii
Samenvatting ................................................................................................. iv
List of abbreviations ....................................................................................... v
Introduction .................................................................................................... 1
Change in consumer behaviour............................................................... 2
Collaboration among courier companies ................................................ 2
Fleet optimization ................................................................................... 3
Chapter 1 : Development Of An Electric Tricycle For Service Companies And
Last-Mile Parcel Delivery............................................................................... 6
1. Introduction ....................................................................................... 6
2. Literature research ............................................................................. 7
2.1. Last-mile delivery trips ............................................................. 7
2.2. Width of the cargo bike ............................................................ 9
2.3. Cargo bike capacity .................................................................. 9
2.4. Electrical assistance ................................................................ 11
3. Summary Of The Research Methodology And Quantified Goals ... 12
4. Conceptual design ........................................................................... 14
5. Stability And Manoeuvrability ........................................................ 16
5.1. Drive system ........................................................................... 16
5.2. Software .................................................................................. 18
5.3. Hinged connection .................................................................. 20
6. Conclusion ....................................................................................... 23
Chapter 2: Handling qualities of a new last-mile vehicle ............................. 24
1. Introduction ..................................................................................... 24
2. New Last-Mile Concept................................................................... 24
3. Benchmark models .......................................................................... 27
4. Experimental Setup ......................................................................... 29
4.1. Measurement procedure .......................................................... 29
4.2. Braking test ............................................................................. 30
viii
4.3. µ-split braking test .................................................................. 32
4.4. Steady-state turn ..................................................................... 33
4.5. Figure of eight ......................................................................... 34
4.6. Prompt steering ....................................................................... 34
4.7. Double lane change ................................................................. 35
4.8. Slalom ..................................................................................... 36
4.9. Summary ................................................................................. 36
5. Results ............................................................................................. 38
5.1. Braking test ............................................................................. 38
5.2. µ-split braking test .................................................................. 39
5.3. Steady-state turn ..................................................................... 40
5.4. Figure of eight ......................................................................... 42
5.5. Prompt steering ....................................................................... 42
5.6. Double lane change ................................................................. 43
5.7. Slalom ..................................................................................... 44
6. Discussion ....................................................................................... 44
7. Conclusion ....................................................................................... 45
Chapter 3: Pedalling comfort of a custom Pedal Series Hybrid Drivetrain in a
cargo e-tricycle ............................................................................................. 47
1. Introduction ..................................................................................... 47
2. Pedal Series Hybrid Drivetrain ........................................................ 47
2.1. Advantages ............................................................................. 48
2.2. Disadvantages ......................................................................... 49
2.3. Examples of PSHD ................................................................. 49
2.4. Pedal Properties ...................................................................... 52
3. Research objectives ......................................................................... 55
4. Tricycle description ......................................................................... 56
4.1. Hardware ................................................................................ 56
4.2. Control loop ............................................................................ 56
5. Methods ........................................................................................... 58
5.1. Pedal generator ....................................................................... 58
ix
5.2. Measurement bicycle .............................................................. 59
5.3. Comparison method ................................................................ 59
6. Results ............................................................................................. 59
7. Discussion ....................................................................................... 61
8. Conclusion ....................................................................................... 62
Chapter 4: Effects of a Torsion Spring Used in a Flexible Delta Tricycle ... 63
1. Introduction ..................................................................................... 63
2. Concept ............................................................................................ 63
3. Tilting/Tipping over ........................................................................ 65
3.1. Bicycle stand ........................................................................... 66
3.2. Locking mechanism ................................................................ 66
3.3. Angle stop ............................................................................... 67
3.4. Torsion system ........................................................................ 68
4. Used system ..................................................................................... 69
4.1. Effects of torsion system on stationary stability ..................... 70
5. Discussion ....................................................................................... 77
6. Conclusion ....................................................................................... 77
Chapter 5: Valorisation plan ......................................................................... 78
1. Market research ............................................................................... 78
1.1. Courier, Express and Parcel services ...................................... 78
1.2. Pickup points .......................................................................... 79
1.3. Other uses ............................................................................... 80
1.4. Bicycle market ........................................................................ 80
1.5. Region ..................................................................................... 80
2. Important design specifications ....................................................... 81
3. Competitors ..................................................................................... 81
4. Product features (unique selling points) .......................................... 85
4.1. Cargo capacity ........................................................................ 85
4.2. Handling qualities ................................................................... 85
4.3. Configuration .......................................................................... 85
4.4. Drivetrain ................................................................................ 85
x
5. Cost estimates .................................................................................. 86
6. Required improvements ................................................................... 86
7. Selling components ......................................................................... 87
Chapter 6: Conclusion and future work ........................................................ 88
Bibliography ................................................................................................. 90
Figures ........................................................................................................ 105
Tables ......................................................................................................... 107
1/107
Introduction
The concept of "last-mile delivery" refers to the final stage in the process of
delivering goods to customers, as defined by the Cambridge Dictionary. It
represents the transition from a transportation hub to the final destination. In
the context of e-commerce, this term describes the last part of the journey
before a parcel reaches the costumer. Although the last mile is just a small
portion of the overall route, the parcel must first reach the transportation hub,
which is known as upstream logistics.
Parcels can travel varying distances, ranging from a few hundred to several
thousand kilometres [1]. With electronics and fashion accounting for 60% of
the total European e-commerce market share [2], and the Asian market
accounting for 52% of the worldwide market share in these sectors, it is
evident that a significant number of parcels are likely to travel several
thousand kilometres. Despite constituting less than 1% of the parcels total
travel distance, research estimates last-mile delivery cost more than half of the
overall logistics cost with estimates reaching 75% [3-6]. This discrepancy can
be attributed to several factors.
Firstly, while upstream logistics primarily involves business-to-business
(B2B) transactions, last-mile delivery is focused on business-to-customer
(B2C) interactions. Consequently, businesses in upstream logistics benefit
from higher parcel volumes, enabling them to combine deliveries. These
combined shipments are often transported in shipping containers via ships,
trains, and trucks. Ships and trains offer the lowest emissions per one-ton mile
during long-haul transportation [7], and even trucks have lower emissions per
one-ton mile compared to vans [8].
Secondly, last-mile delivery involves a higher distance travelled per parcel.
This can be attributed to several factors. As mentioned earlier, last-mile
delivery typically involves individual parcels per delivery address, resulting
in a significantly higher number of stops for the same parcel quantity
compared to upstream logistics. Delivery addresses are often spread across a
region, leading to challenges such as congestion, parking problems, and
inefficient routes. In contrast, upstream logistics experiences fewer traffic
issues as their warehouses are strategically located near major traffic routes.
Thirdly, e-commerce platforms often guarantee next-day or even same-day
delivery, depending on the region. This imposes time constraints and limits
the number of parcels that can be combined in a single trip. In contrast,
upstream logistics faces fewer time constraints, allowing for the consolidation
of deliveries and further optimization.
2/107
These disparities between last-mile and upstream logistics highlight
significant differences and explain the higher cost of last-mile delivery. The
last-mile delivery process is characterized by numerous inefficiencies, each of
which requires unique optimization solutions. However, there is no one-size-
fits-all solution that addresses all these challenges. Outlined below are some
of the solutions identified to improve last-mile delivery:
Change in consumer behaviour
Modifying consumer expectations and behaviour can greatly enhance the
efficiency of last-mile delivery. Encouraging customers to be more patient and
flexible with delivery times allows courier companies to consolidate parcels
within the same geographic area. By increasing the number of parcels
combined in a single trip, the distance travelled per parcel is reduced. This
approach requires effective communication with customers to manage their
delivery expectations and provide incentives for opting for longer delivery
windows.
Furthermore, the adoption of parcel lockers as an alternative delivery method
offers significant advantages. Parcel lockers act as centralized collection
points where customers can retrieve their parcels at their convenience.
Implementing this approach reduces the number of stops required by couriers,
leading to a reduction in trip distances and increased delivery efficiency. It
also provides customers with flexibility and convenience, as they can collect
their parcels at a time that suits them best.
Collaboration among courier companies
Cooperation and collaboration among courier companies can bring about
significant improvements in last-mile delivery. By exchanging parcels or
coordinating their delivery routes, courier companies can optimize their
vehicle routing and achieve greater efficiency. The collaboration can also
involve sharing resources, such as warehouse facilities and sorting centres,
and company specific pickup and drop-off points.
This minimizes duplicated efforts and reduces overall travel distances.
Costumers would receive multiple parcels through one courier instead of
receiving multiple couriers with a single parcel. Furthermore, a shared
network of pick-up points, accessible to customers of multiple courier
services, enhances convenience and reduces the number of dedicated pick-up
locations.
3/107
Fleet optimization
Optimizing the fleet composition used for last-mile delivery is crucial for
efficiency gains. Electric vans present a promising alternative to conventional
vans, offering various advantages. Electric vans have lower operating costs
due to reduced maintenance requirements and lower fuel costs compared to
their fossil fuel counterparts. Moreover, electric vans produce less emissions,
contributing to environmental sustainability. They are also more likely to
comply with low-emission zones, avoiding potential restrictions and penalties.
The adoption of electric vans by courier companies can lead to improved cost
efficiency and reduced environmental impact.
In addition to electric vans, cargo bikes are emerging as a highly viable
solution for last-mile delivery. Cargo bikes offer numerous benefits compared
to traditional vehicles. They have a lower total cost of ownership, as they have
a lower purchase price, require less maintenance and have lower fuel costs, if
any. Cargo bikes are highly manoeuvrable and can navigate more efficiently
through congested urban areas, reducing delivery times and minimizing traffic
disruptions. Their smaller size allows for easier parking and eliminates the
need for large parking spaces or loading docks. Cargo bikes can take
advantage of dedicated bicycle lanes, enabling them to follow direct and
efficient routes that are often unavailable to motorized vehicles. These factors
make cargo bikes an attractive option for last-mile delivery, especially in
dense urban areas.
However, at the start of this research, there is a limited choice in cargo bikes
equipped for courier services. The majority of cargo bikes are consumer-
oriented and lack the necessary features and capabilities to effectively support
the rigorous requirements of courier operations. These cargo bikes typically
have limited payload capacities in terms of weight and volume.
Consequently, there is a clear need for the development of a cargo bike that is
purpose-built for the unique challenges and demands of last-mile delivery in
the courier industry. This PhD research aims to address the aforementioned
challenges by focusing on the development of an optimized cargo bike tailored
specifically for last-mile delivery in parcel couriers.
The first chapter focusses on the comprehensive development of the concept
vehicle designed specifically for last-mile delivery. To obtain a general
overview, numerous stakeholders from diverse fields associated with last-mile
delivery were interviewed. These discussions helped to identify the limitations
of existing cargo bikes and gather the desired specifications for the new
vehicle. The Quality Function Deployment (QFD) method is used to establish
the priorities for the vehicle. By incorporating the insights gained from this
analysis, a cargo bike prototype is developed that is specifically tailored to
4/107
address the unique requirements of parcel couriers. This resulted in a tilting
rear-load cargo tricycle as the optimal solution. A prototype of this tricycle
was constructed and it’s specifications and features are examined. The tilting
mechanism is achieved by a central hinge between the driver and cargo
module. This allows the driver to lean while manoeuvring similar to a
conventional bicycle. Prototype one allowed for modifying the mounted
orientation and height of the tilting mechanism. Multiple configurations were
tested to evaluate the influence on the vehicle handling qualities. An optimal
configuration was gathered and incorporated in the second prototype.
The second chapter offers a comprehensive evaluation of the novel
prototype’s handling qualities. The study aims to objectively evaluate and
asses the vehicle’s performance and behaviour under various loading
conditions and multiple manoeuvres. A comprehensive experimental setup
was designed and executed. The setup is based on existing testing conditions
used in the automotive industry adapted for cargo bicycles. The chapter
provides a detailed explanation of the testing methodology, including the
selection of key performance metrics and the procedure employed to collect
data during the experiments. The results are presented and analysed shedding
light on the vehicle’s handling characteristics. These results are compared
using similar metrics from other cargo bikes. The chapter observes improved
handling qualities compared to the other cargo bikes. Furthermore, chapter 2
discusses any observed limitations or challenges encountered during the
testing process. It highlight the potential areas for improvement, missing
specific type dependent performance metrics and provides insights for future
research and development of last-mile vehicles.
Chapter three delves into a detailed exploration of the Pedal Series Hybrid
Drivetrain (PSHD) that has been developed for the cargo bike. The study
begins by explaining the concept of a digital pedal drivetrain, providing an
overview of the current state of art presenting relevant examples. In the PSHD
system, the driver receives feedback in the form of pedal resistance, which is
proportional to the power required to propel the vehicle. A critical aspect of a
well-designed digital drive is to provide a natural pedalling experience, where
the resistance varies in accordance with the orientation of the pedal crank. This
ensures that the rider experiences a familiar and comfortable pedalling
sensation. The chapter proceeds to explain the digital drive’s control loop
employed in the prototype. The control loop is responsible for adjusting the
resistance based on the rider’s input, the pedal and motors rotational speed and
the desired output. Additionally, the chapter elaborates on the method used to
objectively compare the pedalling feeling to that of a conventional bike. These
methods allow for a comprehensive evaluation of the digital drive. The
findings shed light on the performance of the PSHD and its ability to provide
a pedalling experience that closely resembles a conventional bike.
5/107
Chapter four focusses on the influence of a torsion spring system integrated in
the tilting mechanism of the prototype. As the development and testing of the
first prototype progressed, a crucial issue arose concerning the stability of the
driver's module while parked, leading to the demand for a solution in the
second prototype. In the initial design, the vehicle's unstable equilibrium when
at rest required the use of a separate stand to prevent tipping over. To tackle
this problem, this chapter explores various concepts. Among the proposed
solutions, the metal-elastomer torsion spring system emerges as the most
advantageous and effective solution. This innovative system incorporates an
integrated angle limit, ensuring that the driver’s module remains stable and
avoids tipping over even in extreme conditions. Moreover, the elastomer
torsion spring passively keeps the module upright without any additional
actions from the driver. The chapter presents the calculations and design
considerations that led to the development of an optimal torsion spring
specifically tailored for parking situations, while still having a limited
influence during manoeuvring. By implementing the metal-elastomer torsion
spring system, the tilting delta tricycle demonstrates low-speed stability and
improved parking which are ideal properties for a last-mile vehicle.
Through these investigations, the research seeks to provide insights and
recommendations that contribute to the advancement of sustainable and cost-
effective last-mile logistics practices.
6/107
Chapter 1 : Development Of An Electric Tricycle
For Service Companies And Last-Mile Parcel
Delivery
This chapter is based on the paper “Development Of An Electric Tricycle For
Service Companies And Last-Mile Parcel Delivery” published in the journal
“transport problems” [9].
1. Introduction
The accessibility of urbanised areas for conventional trucks and even vans is
often restricted. Window times, low emission zones, noise limits, ... are some
known restrictions. In a number of historical centres or pedestrian zones, the
access of trucks, vans and also passenger cars is even radically prohibited.
Such bans pose a problem not only for companies that have to deliver e-
commerce parcels but also for service companies that have to deliver spare
parts and tools on site. The regulator points, among other things, to bicycles,
are seen as solutions to the posed problem [10]. Classic pedal-powered
bicycles are suitable for the delivery of light parcels of limited size. The use
of electric bicycles (e-bikes) can increase the comfort of the cyclist and the
area within which the cyclist can make deliveries. The use of bicycles with a
cargo space or a trailer is proposed as a solution for delivering relatively large
parcels and for service companies, among others, by the European
Commission [11] and in recent research results [12-14]. Research and pilot
projects conclude that electric cargo bikes for last-mile delivery improve
efficiency, are less polluting, relieve congestion and reduce costs compared to
vans [15-17]. The COVID-19 pandemic contributed to an expansion in e-
commerce and cargo bike use [18-20]. Professional cargo bike couriers point
out relevant shortcomings of the bikes they use. Companies that lease and/or
maintain cargo bikes for professional users also notice limitations in loading
volume, stability, robustness, ..., especially when compared to classic delivery
vans. In this research an electric cargo bike is developed. On the one hand, it
fits within the mentioned access restrictions in urbanized areas. On the other
hand, it addresses the primary complaints of professionals who cannot use
their classic vans in these areas.
7/107
2. Literature research
Cargo bikes are an optimal alternative for last-mile delivery in urban areas. In
Belgium, 56% of deliveries happen in urban areas [21, 22].
2.1. Last-mile delivery trips
A study conducted in Germany state a maximum daily trip distance of 166 km
for bike and 253 km (van) couriers with an average trip distance of 42 km and
66 km as seen in figure 1. The average single shipment distance is 5.1 km for
bike courier and 11.3 km for car shipments [23]. This is consistent with other
research that concluded an average shipment distance between 2 and 9 km and
an average trip distance of 108.6 km [24].
Figure 1 Bike vs. car shipments: (left) cumulative frequencies of single
shipment distances (right) cumulative frequencies of aggregated daily
mileages [23].
This would account to a range between 6 to 19.6 stops per day. This is however
not in accordance to the majority of sources [25]. This can be partially be
attributed to freelance couriers being included in the dataset. They have a
significantly lower delivery amount due to mainly doing single parcel trips
and having shorter working hours (mean of 5 hours per day).
Furthermore multiple deliveries can be done per stop and not every research
differentiates this. A case study in London noted an average of 37 stops (118
parcels) delivered on a day with a maximum of 72 stops (274 parcels) [26].
This is consistent with a report stating an average of 9 stops per hour which
results in 72 stops per day for an 8 hour shift [27]. Vans have a higher
occurrence of multiple parcels per stop as they have more parking difficulties
8/107
compared to cargo bikes. Courier, Express and Parcel (CEP) services report
200-plus parcel deliveries per day [28].
67% of parcels weigh less than 1 kg and 43% less than 0.5 kg as seen in figure
2.
Figure 2 What was the approximate weight of this particular purchase
[29]?
This concludes in an average weight of 0.9 kg. The average parcel volume
amounts to 13 litres [30].
Using the average weight and volume combined with the 200 daily parcels,
daily volume and weight delivery is calculated. This amounts to a total parcel
volume of 2600 litres with a weight of 180 kg.
As the cargo bike will primarily be used in city centres, it will remain in close
proximity of the distribution centres and is able to reload the cargo and swap
the battery. This means a cargo bike is able to do 2 trips per day. Considering
a maximum of 200 daily stops and one time reload, the cargo bike requires a
minimum payload capacity of 1300 litres, a weight of 90 kg and a range of
80 km. Taking an energy consumption of 2.3 kWh/100 km into account a
battery of at least 1.8 kWh is required [31-33].
9/107
2.2. Width of the cargo bike
The majority of current cargo bikes have a width between 0.8 and 0.9 m [34].
Heavy cargo bikes vary from 0.8-1.2 m. European Regulation limits the
maximum width of a bicycle to 1 m [35]. Individual countries can exclude
those limitations from their own. For example, in Belgium and the Netherlands,
single-track vehicles have a maximum width of 0.75 m, and multitrack
vehicles have a maximum width of 2.5 m and 1.5 m, respectively [36, 37]. The
maximum width of 2.5 m is extremely wide and is stated in regards to
multi-person go-carts or a so-called beer bike. Trailers have a maximum
allowed width of 1 m in general and 1.2 m for permitted pilot projects [38].
To the best of the authors knowledge, there is no legislation in place in
Belgium regarding the width of bicycle lanes. There are however
recommendations stating a minimum width of 1.5 m, taking a 1-m-wide
bicycle into account with 0.25 m clearance on both sides [39]. This is currently
not the case for all bicycle lanes. Currently, 42% of cyclists are not satisfied
with the width of Belgian bicycle lanes, and 31% of current cycle lanes are
narrower than the prescribed 1.5 m [40]. Denmark recommends a width of
2.2 m [41]. Greibe et al. recommend a minimum width of 1.95 m for a two-
lane cycle track [42]. Cargo bikes with a maximum width of 1 m are allowed
to use bicycle lanes [43]. A cargo bike wider than 1 m cannot benefit from the
advantages of using bicycle lanes. Belgium, Denmark and the Netherlands are
described here since they are part of the top five in terms of Europe’s main
cargo bikes market [44] and are the top three in private cycling use [45]. Due
to financial and political incentives, Belgium emerged as a key cargo bike
market in 2021 [46].
2.3. Cargo bike capacity
Figure 3 shows different types of cargo bikes with their respective payload
information and widths [47]. Figure 4 shows the representation of the types.
Overall, 51% of the cargo bike sales are three-wheeled cargo bikes, and 49%
are two-wheeled [48].
Figure 3 Cargo bike types [13].
10/107
Figure 4 Representation of cargo bike types [49].
In general, the most common cargo bicycles are front-load cargo bikes, as 62%
of cargo bike trips are carried out by front-load cargo bikes [49]. Messenger
bikes have a limited cargo capacity but are commonly used for food delivery
since they require a low cargo capacity, and due to delivery within one hour,
deliveries cannot be combined. Postal services also use this type of bicycle for
letter deliveries.
Front-load cargo bikes are the most common cargo bikes for private use. They
can transport up to 125 kg, which is perfect for grocery shopping or
transporting children. This type is also often used by parcel couriers. The
cargo bike has a decent capacity and has similar handling qualities as a
traditional bicycle. However, during interviews it was noted the front-load
bikes still have a limited volume capacity and a trailer is often used to increase
the capacity. The full weight capacity is rarely reached due to limited volume
and low-density parcels. Research also points to limited capacity as a major
shortcoming [50, 51]. Current market trends show an increase in cargo
capacity [52]. Rear-load cargo trikes in general have the highest cargo
capacity. The extra wheel distributes the load and allows additional cargo
capacity. Placing the cargo behind the driver allows for a higher cargo area
without obstructing the driver’s view. These types are often used by parcel
couriers such as UPS, DHL and FedEx [53-55]. However, current rear-load
cargo trikes increase the average travel time [56]. This is stated to be due to
the extra cargo they can carry.
The specifications in figure 3 are the average values for these cargo types.
There are extreme versions, such as the Cargo Bike XXL and the Urban Arrow
Tender as seen in figure 5.
11/107
Figure 5 Ziegler Cargo Bike XXL [57] and Urban Arrow Tender [58].
The Cargo Bike XXL from Ziegler is capable of delivering up to 500 kg with
a maximum volume of 4.3 m³. Its total length is 6.5 m. Due to the large loading
capacity, the cargo bike needs an extra-wide cycle infrastructure. The Urban
Arrow Tender is a tadpole tricycle with a payload capacity of 300 kg and
1.2 m³. The vehicle is 1.14 m wide.
2.4. Electrical assistance
E-bikes assist the driver and make it easier to transport high payloads. Electric
assistance generally reduces delivery times, especially along routes with
elevation differences [56]. Currently, 40% of cargo bikes are electrically
assisted. In 2020, 92% and 75% of all sold cargo bikes were electric in Europe
and Germany, respectively, with the Electric Pedal Assisted Cycle (EPAC)
representing the majority of cases [46, 59]. EPAC is currently limited to
250 W and 25 km/h. These electric cargo bikes are the same type as e-bikes.
This, however, provides limited assistance when cycling with a high payload
on hilly terrain. A different category with an assistance motor power of
1000 W can also be used. This category fall under L1e-A and requires type-
approval in Europe which reduces the development towards this type.
Furthermore, the majority of European countries oblige wearing a helmet,
insurance and a licence to drive this type [36].
12/107
3. Summary Of The Research Methodology And Quantified
Goals
In this research, a systematic scientific method has been applied for the
development of the cargo bike, namely the Quality Function Deployment
(QFD) method [60]. To obtain a full view of the desired features of the vehicle
to be developed, representatives of all types of stakeholders were interviewed.
In total more than 30 interviews were conducted with representatives as
summarized below:
1. experienced bike couriers: couriers transporting light parcels of
limited size on a conventional bicycle, couriers using a two-wheel
cargo bike, couriers using a tricycle or quadricycle, couriers using
a trailer
2. operators of a fleet of bicycles and tricycles, including cargo bikes
3. cargo bike distributors
4. manufacturers of cargo bikes
5. those responsible for mobility and logistics at local authorities
6. operators of an urban distribution centre
7. classic courier services using mostly or exclusively vans and trucks
8. service companies located in an urban environment with access
restrictions
9. service companies located in rural areas but with customers in
urbanised areas
10. shopkeepers in urbanised areas
The desired features of a cargo bike detected during the interviews are
summarised as follows:
1. highly manoeuvrable
2. robust in the sense of being able to withstand daily use by non-
owners of the bike
3. allowed by legislation on bicycle paths
4. low wind resistance
5. loading floor at least suitable for a euro pallet (80 x 120 cm)
6. able to transport long pieces
7. stable at all times, even with loads of 150 or 200 kg
8. can be parked without external support
9. secured against theft
10. safe at all times, even when driving downhill fully loaded
11. battery rechargeable in 15 minutes or replaceable
12. complies with relevant legislation
13. easily repairable
14. intuitive to use
15. cheap to build in small series
13/107
The abovementioned features desired by stakeholders are put in the house of
quality matrix according to the QFD method. The matrix structures the desired
features and distinguishes between requirements and wishes; if the cargo bike
under development does not comply with the requirements, it will not be
accepted by the market. The wishes are rather desirable features. According
to the QFD method, the final classification as “requirement” or “wish” is done
by the team of developers using the details of the wording by the interviewees.
In this phase of the development – which took place in 2019 – over 300 types
of existing cargo bikes are checked with the stakeholder requirements, and it
is concluded that not a single type of existing cargo bike complies with all
requirements. Therefore, the development process is continued by creating a
list of engineering requirements. The engineering requirements aim to
quantify the desired features obtained through the interviews:
1. the cyclist has a free view of the road
2. maximum total load (cargo + cyclist) of 275 kg or more
3. maximum cargo payload of 200 kg or more
4. maximum total payload of each wheel of 110 kg or more
5. maintenance free for 1000 driving hours at maximum load
6. maximum width of 1 m
7. s-turn of two times 180° and a radius of 4 m at a speed of 15 km/h
or more
8. all wheels diameter of 20" or more
9. minimum ground clearance under the loading floor of 14 cm or
more
10. stability is an intrinsic property of the bike
11. loading floor of 81 x 121 cm or more, with a pallet edge of 5-15 cm
high on four sides
12. possibility of mounting a ladder-shaped front wall on the loading
floor for transporting long pieces protruding above the cyclist
13. assistance speed is limited to 25 km/h
14. can be homologated as an L1e-A or L1e-B vehicle [61], or the
maximum rated output power of the motors is limited to 250 W.
15. replaceable battery with a capacity of at least 2000 Wh and space
for a second identical battery
16. parking using a mechanical brake or active position control using
the motor
17. automatic support leg in case of a two-wheel bike, slide in and out
within 0.3 s
18. hidden built-in tracing system operational over 200 km
19. automatic quality lighting with EN number
20. conventional brake levers for bicycles (on the right-hand side for
rear brake control, on the left-hand side for front brake control)
21. proportional regenerative brakes operated by a conventional brake
lever
14/107
22. braking deceleration better than 4.4 m/s² on (clean) asphalt or
concrete roads [62]
23. assembly and maintenance can be done using readily available tools
only
24. use of parts that are commercially available except for the frame
25. tuned to a cyclist of 1.8 m and adaptable for cyclists between
1.7 and 1.9 m
The above list of engineering requirements, referred to as ER1 to ER25, is to
be considered as a checklist during the next development phases, namely the
creative conceptual design and detailed engineering phases.
Multiple concepts for new cargo bikes were designed based on previous
requirements. Their potential to meet the engineering requirements has been
evaluated, and the most promising concept has been selected. It is
schematically represented in figure 6.
4. Conceptual design
Figure 6 Concept of the new type of cargo bike.
This cargo bike can be classified as a rear-load cargo tricycle. A rear-load
cargo tricycle typically allows a free view on the road (ER1) independent of
the dimensions of the cargo. A tricycle is easier to keep stable than a bicycle
and can be suitably dimensioned for cargo of a euro pallet size. The specific
proposed design consists of a part for the cyclist at the front (cyclist module)
15/107
and a part for the cargo at the back (cargo module). Both modules are
connected by a hinge.
The cyclist module consists mainly of a front wheel with a fork and
handlebars, a saddle, a pair of pedals mounted on an electric mid-drive motor
and a frame. For the front wheel, a conventional 26" or 28" spoked wheel is
chosen, of which numerous variants are commercially available with matching
tires (ER24). The fork and handlebars are also parts for which numerous
variants are commercially offered. The saddle is longitudinal and height
adjustable (ER25). The frame is a conventionally welded tubular frame
commonly used in cargo bikes. The pedals and handlebars are conventionally
commercially available components. The pedal rods are mounted on a
conventional mid-drive motor for an e-bike. However, this motor is used as a
generator by installing purpose made software on the motor controller. The
front wheel is equipped with a conventional hydraulic brake to meet legal and
engineering requirements. This front wheel brake is considered an emergency
brake and should not be used often, as it is the most wear-sensitive part of the
tricycle.
The cargo module consists mainly of a U-shaped frame. A wheel motor is
mounted on each leg of the U, and the loading floor is situated between both
legs. The width of each leg is limited to 9.5 cm to allow for a width of 81 cm
for the loading floor (ER11) within the maximum total width of 100 cm (ER6).
The ground clearance under the loading floor can be chosen to comply with
the engineering requirements (ER9) since it is not determined by the diameter
of the wheels. The U-frame is composed of sheet metal parts connected by
structural rivets (ER23). This construction method avoids the use of moulds
and dies, allows the strength (ER2-ER3) and stiffness to be maximised by
means of computer simulations, avoids the use of specifically trained
personnel for welding (ER23) and ensures fast assembly. Furthermore, well-
shielded compartments for electronics (ER18), batteries (ER15) and cables
can be integrated inside the U-shaped frame owing to the sheet metal-based
concept. A ladder-shaped front wall (ER12) and high-quality lighting (ER19)
are also included.
The cyclist module and the cargo module are hinged so that the cyclist module
can rotate around an almost horizontal longitudinal axis in relation to the cargo
module. This enables the driver to lean when taking corners, enhancing the
stability of the tricycle (ER10). The exact choice, location and orientation of
the hinge have been experimentally researched and are discussed below.
The loading floor is surrounded on three sides by the U-shaped frame. At the
backside, there is a robust door that can be folded down around horizontal
hinges. When folded down, this door supports the loading module to prevent
16/107
the tricycle from tipping over backwards when loading. Loading can be done
from above with an overhead crane, from behind with a pallet truck or
completely manually from the side. Most likely, a service company will mount
a fixed box of suitable height on the loading floor. Couriers are expected to
prefer to load a single case or two stacked cases containing the parcels to be
delivered. Foldable cases can be used to minimise wind resistance.
All engineering requirements not yet mentioned in this paragraph – except for
the manoeuvrability (as quantified in ER7) and stability requirements – can be
dealt with in the detailed engineering phase. Further research was focused on
these features, as they are crucial for couriers in urban areas.
5. Stability And Manoeuvrability
As the main dimensions of the tricycle are determined by the need to transport
cargo the size of a euro pallet, the stability and manoeuvrability depend mainly
on the complex mechanical drive system, the control software and the proper
features of the hinged connection between both modules.
5.1. Drive system
The propulsion system of the proposed tricycle consists of a wheel motor in
each of the rear wheels, a generator mounted between the pedal rods, one or
more battery packs, motor control electronics with the necessary software and
controls on the handlebars. The wheel motors can provide torque for driving
or braking and a holding torque when parked.
A permanent magnet brushless DC (BLDC) motor with coils on the stator and
magnets on the rotor has been chosen as a wheel motor. Two flanges are
mounted on the rotor. They support the wheel rim on which pneumatic tires
are mounted. 20" tires (ER8) with a payload of 140 kg are commercially
available (ER4). Three hall sensors are present between the stator coils used
for the motor control and to determine the motors rotational velocity. The
power cable and signal cable are routed outward along the axis. The stator can
be powered by a 24V, 36V or 48V DC power source according to the desired
maximum speed. A direct drive BLDC motor is chosen instead of the more
widely available geared motor because such a direct drive motor can be
adapted to the very small width that is available. The direct drive motor,
including a stator shaft, rotor flanges and a rim, forms a single unit with a
drop-out size of 70 mm. It can be fitted within the width of 95 mm of the leg
of the U-shaped frame. The wheel motors propel and stop the vehicle using
regenerative braking, increasing the range as recommended by previous
research [63].
A mid-drive motor of a conventional e-bike is mounted on the cyclist module.
It is powered by the cyclist using the pedals, and it serves as a generator. It is
17/107
connected to the battery pack that powers the wheel motors, thus increasing
the range by the energy supplied by the cyclist. This type of series hybrid
system was studied in detail by [64, 65].
Numerous user control features are implemented to enable the smooth
operation of this drive system:
1. an on/off key interrupting the connection to the main power source;
only the tracer (ER19) remains powered
2. an “intention” multi-position switch:
• Parking: the wheel motors are energised and controlled to a
position. This prevents the tricycle from spontaneously moving
on a slope (ER16)
• Driving: the wheel motors are energised depending on the
movement of the pedals. Up to a speed of 5 km/h, the wheel
motors can be driven proportional to the position of a throttle,
serving as a walking function
• Reverse: similar to the walking function but backwards
3. an “assistance” multi-position switch to select the level of
assistance by the tricycle rear wheel motors. The switch position
mainly determines a multiplication factor to calculate the torque of
the motors, starting from the torque of the cyclist on the pedal rods
mounted on the generator
4. a “throttle” to control the power of the wheel motors up to 5 km/h
forwards and backwards
5. a brake lever on the right side of the handlebars that, when operated,
immediately interrupts the propulsion by the wheel motors, causing
the tricycle to coast. When the brake lever is pressed further, the
regenerative braking action of the motors is activated. The position
of the brake lever determines the power in braking mode (ER21)
6. a hydraulic brake on the front wheel of the tricycle (ER20). It is
activated by a brake lever on the left side of the handlebars. A brake
switch is integrated into this brake lever. When operated, it
immediately interrupts the propulsion by the wheel motors and sets
a fixed braking current. When the brake lever on the right-hand side
is also pressed, this brake lever determines the power in braking
mode. The hydraulic brake can also be used as a parking brake on
the front wheel
7. a display providing information to the cyclist, such as speed,
distance travelled, battery charge, ...
8. a brake resistor mounted in the frame. It dissipates energy from the
generator when the batteries are fully charged and limits the energy
to the battery from regenerative braking in order to maintain the
health of the battery
9. an electronics box, which includes a double driver for both wheel
motors, a driver for the generator, a data processor and a data
18/107
storage module. The data storage function was used intensively to
gather data for the ongoing research. Performance data can be
visualised on the display, while more detailed and historical data
can be visualised on a smartphone over a link such as Bluetooth
The handlebar of the tricycle is shown in figure 7 below.
Figure 7 Handlebar of the tricycle.
5.2. Software
The software used to run the tricycle mainly consists of modules for the wheel
drives, the generator, battery management and user interaction. This research
focused on the software for the wheel drives and the generator.
The drive of the wheel motors controls the torque exerted on the rotor of each
motor. When driving straight ahead with equal load and tire pressure on both
wheels, this torque should be identical for both motors. As soon as the
handlebars are turned when entering a bend, the resistance on the wheel on the
inside of the bend increases, and the resistance on the wheel on the outside of
the bend decreases. The speed of the motor on the outside of the bend increases
while the opposite occurs for the motor on the inside of the bend, thus helping
the bicycle turn corners. The individual control of both wheel motors increases
the manoeuvrability of the tricycle considerably in a way comparable to the
Electric Stability Control (ESC) implemented in conventional cars.
Furthermore, this drive reads the inputs of the intention switch, assistance
switch, throttle and both brake levers mounted on the handlebars. As
mentioned before, the brake lever on the right-hand side is used to control the
level of regenerative braking by both wheel motors. Apart from the advantage
of energy recuperation, using the wheel motors as primary braking also
activates a software-implemented Anti-lock Braking System (ABS) without
requiring extra components. The rotational speed of the wheel motors is
constantly monitored using three Hall sensors. A measured decrease in the
rotational speed of a wheel motor is interpreted by the software to identify the
slipping of the wheel. In that case, regenerative braking is interrupted
Electric brake
On/Off key
hydraulic brake
Intention
switch
Assistance
switch
throttle
display
19/107
immediately until the wheel regains traction. Then, braking is reactivated as
before. This procedure is repeated as long as the wheel tends to slip. As a
consequence, the maximum vehicle deceleration mainly depends on the
coefficient of friction between the tires and the road. This coefficient is beyond
0.5 anyway, thus allowing a deceleration above 4.4 m/s² (ER22).
This drive also determines the maximum speed at which the motors still
provide assistance. This switch-off speed depends on several parameters.
Local legislation can limit this switch-off speed (often to 25 km/h). In practice,
the voltage supplied by the battery varies depending on the State Of Charge
(SOC). For a nominal voltage of 48 V – as used in the tricycle described in the
current work – the actual voltage varies between 39 V (0% SOC) and 54.6 V
(100% SOC). Battery life can be prolonged by keeping the SOC between 20%
and 80% [66]. Other research [67] shows that the efficiency of a BLDC motor
may diminish consistently when operated in the top 10% of its rotational
speed, thus draining the battery very quickly. Therefore, the motor’s switch-
off speed is made variable depending on the actual voltage of the battery. The
switch-off speed is within the limit implied by legislation as long as the battery
voltage is more than 42 V (ER13). As soon as the battery voltage drops below
this value, the switch-off speed is reduced to avoid operation in the inefficient
speed zone. The software of this drive can also limit the output power of the
wheel motors. According to European regulations, the rated nominal
continuous power output of the propulsion of an e-bike – either a bicycle or
tricycle – is limited to 250 W [61]. If not, homologation as a light electric
vehicle category L1e-A or L1e-B is required. The technical guidelines state
that the power measurement may be carried out for half an hour, which means
that the peak power consumption can be significantly higher than indicated in
these guidelines. In concrete terms, this means that the power consumption
when driving up a slope can be significantly higher than 250 W as long as it
is largely reduced when driving down the slope immediately afterwards. The
software manages this power consumption, taking into account energy
recuperation during braking on some downhill roads.
The drive of the generator creates resistive torque on the bottom bracket. A
safe and comfortable feeling during cycling is related to the course of this
torque as a function of the position of the pedal rods. Prior tests reveal that a
constant resistive torque is very uncomfortable for the cyclist [68]; in the dead
points at the upper and lower positions of the pedal rods, a resistive torque
from the generator is very difficult to overcome. In conventional bicycles
these dead points are bridged by the inertia of the moving bicycle. This is not
the case for a tricycle carrying a heavy load starting at a standstill. Further, at
higher speeds, the resistance must be limited in the dead points to maintain
safety. Preferably, the resistance on the pedals also depends on the physical
resistance of the loaded tricycle (i.e. the rolling and air resistance). Since there
20/107
is no mechanical linkage to the rear wheel motors, a digital linkage is
established to create the feeling of the physical resistance of the tricycle. The
actual electrical current and the actual rotational speed of the wheel motors
are indicative of this physical resistance. Therefore, the generator drive uses
this current and speed as input values. The resistive torque on the bottom
bracket created by the generator depends upon several factors:
• the rotational speed of the generator
• the rotational speed of the wheel motors
• the actual current of the wheel motors
• the actual position of the pedal rods
Moreover, this digital linkage serves as a continuous variable transmission
(CVT): a constant pedalling speed is created to maximise the comfort of the
cyclist. The generator control loop is further explained in chapter 3.
5.3. Hinged connection
The hinge consists of a shaft that rotates in a housing. It resists forces in all
directions perpendicular to the shaft, forces in both senses parallel to the shaft
and bending moments in both senses around all transverse directions. The
longitudinal direction of the shaft largely coincides with the longitudinal
direction of the tricycle. However, it can be tilted slightly up or down at the
front to improve manoeuvrability. Moreover, the hinge can be mounted at a
higher or lower position between both modules of the tricycle. A design of
experiments has been set up to determine the ideal tilt angle and installation
height.
The prototype tricycle used for these experiments is shown in figure 8.
Numerous geometrical parameters of this tricycle influence its
manoeuvrability and stability. Since the design of experiments focuses on the
height and tilt angle of the hinge, these geometric parameters are copied from
a rigid cargo bicycle. The design of experiments includes three different
heights (220 mm, 360 mm and 500 mm above ground level) at three different
tilt angles (10° upwards, horizontal and 10° downwards) with an empty load
and the maximum load on the loading floor.
21/107
Figure 8 Prototype with adjustable hinge.
Five cyclists conducted an obstacle course at different speeds (below 10 km/h,
10-15 km/h, 15-20 km/h and above 20 km/h). The specified combinations of
heights and tilt angles were evaluated by questioning the cyclists. They were
explained that “stable” is to be understood as not likely to fall or tip over, and
“manoeuvrable” is to be understood as easy to move and direct as intended.
Parallel to the interviews, the stability was evaluated by measuring the steering
angle. Frequent small changes in the steering angle indicate poor stability of
the cyclist module and, thus, the overall stability of the tricycle.
The experiments show that the height of the hinge position has no impact on
the tricycle stability at low speeds. At higher speeds, cyclists reported feelings
of instability when entering a corner with the hinge in the highest position.
They felt the cyclist module being swept out from under them when tilting the
cyclist module to the inside of the turn. As such, the interviewed cyclists
unanimously preferred the lowest hinge position.
The orientation of the hinge significantly influences both the stability and the
manoeuvrability of the tricycle. Orienting the hinge perfectly horizontally
creates a feeling similar to cycling with a conventional two-wheeled bike. By
tilting the front of the hinge slightly downwards, the tricycle is less stable but
more manoeuvrable. Cornering with a downwards tilted hinge causes a small
rotation of the load module towards the outside bend when the cycle module
is tilted inwards. This rotates the rear wheels away from the direction the
cyclist module is going, creating a smaller turning radius compared to a
horizontal orientation of the hinge. The cases of tilted downwards, horizontal
and tilted upwards hinge are illustrated in figure 9. Unfortunately, a
downwards tilted hinge steers the tricycle sideways. This is experienced as
22/107
instability and causes the cyclist to react. A new user will not expect such
instability and, therefore, will feel uncomfortable. After using the tricycle for
a couple of hours, this uncomfortable feeling seems to disappear. Orienting
the hinge in a slightly upwards position has the opposite effects on
manoeuvrability and stability. The cargo module moves in the same direction
the cyclist wants to turn, thus enlarging the turning radius and making it more
difficult to turn.
Figure 9 Turning radius of the vehicle with a) a downwards tilted hinge,
b) a horizontal hinge and c) an upwards tilted hinge.
These effects are proportional to the tilting angle of the hinge, both for
downwards and upwards tilting. A larger inclination has a larger effect on the
manoeuvrability and on the stability. Finally, the cyclists preferred a slightly
downwards tilted hinge, as the improved ability to make sharp turns outweighs
the mild reduction in stability. Probably, the cyclists also took into account
that they were used to this instability at the end of the series of tests.
Finally, the hinged tricycle with the hinge in the lowest position and tilted
downwards is compared to a fixed frame tricycle. For cycling speeds below
10-15 km/h, the fixed frame tricycle is more stable than the hinged tricycle.
The latter requires more steering corrections. For cycling speeds beyond 10-
15 km/h, the hinged tricycle is more manoeuvrable than a tricycle with a fixed
frame due to the ability of the cyclist to lean into the corners. Stability is
comparable for both hinged and fixed frames with an empty load. When
transporting goods, a fixed frame tricycle can even tilt over in sharp turns,
while a hinged tricycle will show stable behaviour.
a)
b)
c)
23/107
The poor stability of the hinged tricycle at low speeds can be eliminated by
introducing rotational stiffness into the hinge. This stiffness reduces the
tendency of the cyclist to tilt at low speeds. The optimal rotational stiffness
depends on the mass and the centre of gravity of the cyclist. Therefore, an
adjustable rotational stiffness is recommended This is further discussed in
chapter 4.
Lastly, the hinge system also includes a couple of stabilisers that keep the
cycling module in a vertical position when parked. The actual hinge and the
stabilisers are integrated into the frame to avoid damage by water infiltration.
6. Conclusion
Existing cargo bikes do not completely meet the expectations of professional
users for the last-mile delivery of e-commerce parcels, nor do they meet the
requirements of service companies. A novel future-generation tricycle was
developed using the QFD method to inventory the requirements of such a
cargo bike and to select the most appropriate design concept. The systematic
use of design of experiments during the detailed engineering step of the
development resulted in an electric cargo tricycle that is very stable and
manoeuvrable. The vehicle is able to carry a euro pallet and has a payload
capacity of 200 kg. The series hybrid drivetrain allows a hinged frame while
maintaining a low centre of gravity and remaining within width restrictions.
The hinge between the driver module and cargo module allows for a nimble
and comfortable ride. It complies with all important requirements of all
stakeholders detected during the research. The concept corresponds with
current trends in the cargo bike market, as seen in Cargo Chariot and Fulpra.
A prototype of this tricycle is shown in figure 10.
Figure 10 Last-Mile Vehicle prototype.
24/107
Chapter 2: Handling qualities of a new last-mile
vehicle
This chapter is based on the paper “Handling qualities of a new last-mile
vehicle” published in the journal “Journal of Transportation Technologies”
[69].
1. Introduction
Chapter one presented the development of a novel Last-Mile Vehicle (LMV)
that successfully meets the payload capacity and engineering requirements.
For a courier vehicle, impeccable handling qualities are of the utmost
importance. The cargo bike must remain stable regardless of the payload.
The cargo bike must exhibit stability regardless of the payload while retaining
agility navigating through tight spaces, especially in historical narrow city
centres, where it finds optimal utility. Moreover, controllability plays a vital
role in ensuring the vehicle’s overall safety during operation.
In the automotive sector, standardized tests have been established to asses and
measure specific aspects of handling qualities for conventional vehicles.
However, when it comes to bicycles and cargo bikes, such standardized tests
are notably absent. While there are standardized tests for assessing braking
performance, evaluating the overall handling qualities of these specialized
vehicles remains a gap in the industry.
This chapter aims to address this significant issue by delving into the
evaluation and measurement of handling qualities for the LMV. By
developing and implementing relevant testing methodologies, the chapter
seeks to establish a robust framework for assessing the handling
characteristics of cargo bikes and bicycles, filling a crucial void in the field of
last-mile transportation.
2. New Last-Mile Concept
The LMV is a novel concept developed for last-mile delivery as seen in figure
11. The vehicle is designed to transport euro pallet sized cargo with a
maximum load of 250 kg. Being able to transport a pre-packaged city
container or euro pallet smoothens the transition from a transportation hub to
the final destination.
25/107
To ensure that the driver does not have an obstructed view, the cargo bed is
located behind the driver. The euro pallet sized cargo bed is located between
the rear-wheels to ensure a low centre of gravity. It has a load capacity of
around 200 kg. The vehicle consists of 2 modules. A rear cargo module and a
front cyclist module. The prototype is powered using two BLDC hub rear-
wheels with a rated power of 2000 W and a maximum torque of 90 Nm. The
eventual cargo bike requires a total rated power below 1000 W to allow
homologation within L1e-A category or below 250 W for the EPAC-category.
The motors have an independent current control creating an electrical
differential. This allows natural cornering and doesn’t steer the vehicle in one
direction which is often the case with 1WD delta tricycles. The front cyclist
module is connected with a hinge enabling the cyclist to lean during corners.
This creates a natural cycling feeling improving the manoeuvrability of the
LMV.
Figure 11 LMV concept.
Hinging the bicycle module and the cargo module allows the driver to lean
during manoeuvres without noticeable influence from the load. This hinge is
constructed using a slewing ring bearing allowing only one rotational degree
of freedom. The cargo module will not lean during cornering. The position
and orientation are determined to obtain a proper balance between stability
Cargo
module
Cargo
module
Drivers
module
Pedal
generator
Hinge
26/107
and manoeuvrability. Bicycles become more difficult to balance at lower
velocities [70]. To improve the stability during low velocities a torsion spring
is added. A torsion spring consisting of an elastomer structure is added. This
combines torsion and damping improving the stability during low velocities.
Due to regulatory constraints [35], the vehicle has a maximum width of
1 meter. The width of a euro pallet of 0.8 m remains less than 100 mm
available wheel dropout. This results that there is no room available for the
mechanical brakes. That’s why the cargo bike uses regenerative braking. The
front-wheel has an additional mechanical brake. Regenerative braking is
controlled by a brake lever which proportionally regulates the braking of each
motor. While braking, each motor recharges the vehicle’s battery. Li-ion
batteries have limited energy they can safely absorb. The excess braking
energy is consumed using a heat sink to avoid damaging the battery. For extra
road grip, the wheel motors have individual software implemented ABS
optimizing limiting the slip improving the safety of the vehicle.
The flexible construction between the modules prevents a traditional bicycle
drive train. Tilting would misalign the chain and cause a derailment. Another
flexible mechanical coupling is an alternative that would make it more
complex, expensive and would require a higher wheel dropout. An electrical
drivetrain doesn’t require extra installation room and has fewer design
constraints.
The electrical drivetrain consists of 2 wheel motors to propel and a mid-motor
used as a pedal generator. The pedal generator drive controls the pedal
resistance simulating a natural pedal feeling. The force applied on the pedal
generator combined with a proportional rotational difference controls the
wheel motors. The direct-drive hub motors don’t have an integrated freewheel
so it allows control in both directions. This allows the LMV to have a forward
and backward walking support function.
A prototype was constructed to assess the optimal position and orientation to
obtain an optimal balance between stability and manoeuvrability. The second
prototype shown in figure 12 is used to perform the manoeuvres. Since the
bicycle is developed as a cargo bike where the load changes constantly, tests
are conducted with different loads. Drivers have extensively tested the vehicle
before the manoeuvres are performed to get used to the vehicle and get proper
results.
27/107
Figure 12 LMV prototype.
3. Benchmark models
The LMV is compared with QuadRad, City Shopping, Babboe Big and
Babboe City (figure 13). These cargo bikes are a good representation of the
different types of cargo bikes. A four-wheeled bike, a three-wheeled bike with
two rear wheels, a three-wheeled bike with two front wheels and a long John
(two-wheeled bicycle with the cargo hold in front of the driver). These
configurations represent different categories of cargo bikes as seen in figure 3
in chapter 1.
Figure 13 QuadRad, City Shopping Bike, Babboe Big, Babboe City
(from left to right) [71].
The QuadRad is a four-wheeled bicycle with pedal assistance. The concept
vehicle can carry up to 180 kg behind the driver and is designed for
commercial applications and everyday use. The City Shopping Bike has a
fixed delta frame with 2 rear wheels and 1 front wheel. With a revenue load
of 22 kg, the City Shopping Bike is developed for everyday life use. The
Babboe Big is a three-wheeled cargo bike with 2 front wheels and one rear
wheel. This is a tadpole design. The Babboe Big is designed to transport
28/107
children safely on a bench, which is equipped with a seat belt. It can also be
used to transport cargo up to a load of 100 kg. The Babboe City is a Long John
design. It is a two-wheeled bicycle with a cargo area between the driver and
the front wheel. The Babboe City can carry 80 kg. The LMV has a delta design
with a flexible connection between the cargo module on the rear and the driver
module on the front. The LMV is developed for commercial use and can carry
200 kg of cargo. All values are summarized in table 1.
Table 1 Technical specifications
QuadRad
City
Shopping
Babboe
Big
Babboe
City
LMV
Wheel base
(mm)
1350 1320 1360 2000 2000
Wheel track
(mm)
730 635 725 / 910
Tyre size
(in.)
26 26
26 Rear
20 Front
26 Rear
20 Front
20 Rear
26 Front
Chassis
clearance
(mm)
153.5 140 75 150 150
Drive
concept
Mid-Engine Mechanical Mechanical Mechanical
Series
Hybrid
Weight of
vehicle (kg)
60 32.5 61 45.5 122.2
Revenu load
(kg)
180 22 100 80 200
Axle weight
(kg)
FA RA
24 36
FA RA
9.5 23
FA RA
48 13
FA RA
22 23.5
FA RA
30.8 91.4
turning
clearance
circle (m)
4.35 2.92 6.85 5.88 3.80
City Shopping, Babboe Big and Babboe Big are mechanically driven.
QuadRad is mechanically driven with extra assistance from a mid-drive motor
of 250 W and a maximum torque of 70 Nm. The LMV is powered by 2 hub
motors with a rated power of 2000 W and a maximum torque of 90 Nm each.
The power of the LMV prototype exceeds the permitted 250 W for an e-bike.
The LMV can be software limited so the nominal continuous power remains
below 250 W. To classify the LMV as an EPAC different hub motors must be
installed.
The Babboe City and LMV benefit from a low centre of gravity since the cargo
area is positioned between the wheels. This lowers the centre of gravity and
increases stability. Due to the delta design, the City Shopping has the lowest
turning circle followed by the LMV due to the larger wheelbase. Most vehicles
have a higher load on the rear axle. The Babboe Big is the only one with a
higher weight distribution due to the tadpole design. The LMV can carry the
highest cargo load.
29/107
4. Experimental Setup
Handling qualities of a vehicle entail multiple aspects of how the vehicle
handles while driving. This is a balance between the vehicle’s stability and
manoeuvrability.
Aspects like the braking characteristics are important for the overall safety.
Compliance with established safety standards is required to guarantee the
vehicle’s safety [72]. However, evaluating the stability and manoeuvrability
is challenging due to the lack of standardised test for bicycles. The automotive
industry uses standardized manoeuvres to assess the vehicle dynamics of a
vehicle [73]. These tests are open-loop control, meaning the result is
independent of the driver. These tests are used to determine the vehicle’s
safety rating. There are also no standards to asses a motorcycle’s dynamic
behaviour to the author’s knowledge. Motorcycles have a series of tests but
this is to evaluate the driver's skills.
The bicycle industry doesn’t have standard tests to evaluate the handling
qualities of the vehicle. This is due to the tinkering method and subjective
evaluation which has been used for the past 200 years during the evolution of
bicycles [74]. Furthermore, cycling doesn’t require a driving licence which
evaluates driving skills. Kooijman et al. summarize experiments used to
evaluate different aspects of the handling qualities [75]. The manoeuvres
performed as described in [71] are derived from the automotive industry test
and adjusted to the bicycle industry. These manoeuvres allow an objective
evaluation of the different vehicles.
4.1. Measurement procedure
The different manoeuvres are described according to figure 14. The
manoeuvres are set up using pylons.
Figure 14 Legend of manoeuvre setup.
The manoeuvres are not always performed at the desired speed. A difference
in the performing speed changes the results and doesn’t allow for proper
comparison. To compensate for the speed variations, the measurements are
corrected using equation 1 [72].
30/107
=
(1)
With:
=
=
= /
= /
This compensation is applied in each manoeuvre using a specified velocity.
4.2. Braking test
To ensure the safety of the vehicle, the braking capabilities requires
evaluation. When any obstacles occur, the vehicle must be able to perform a
controlled emergency stop. This test measures the braking capabilities at a
given speed. An emergency stop is done at a velocity of 25 km/h. Braking
is applied when the front wheel is between the pylons as shown in figure 15.
Figure 15 Emergency brake.
The test is carried out by applying the front-wheel brake, rear-wheel brake
separately and combined. The emergency brake requires the vehicle to a full
stop. The emergency brake is evaluated by measuring the distance from the
pylons to the front wheel (stopping distance). The braking performance is also
evaluated using the Mean Fully Developed Deceleration (MFDD) [62].
MFDD is calculated using equation 2.
=
.() (2)
With:
= /²
31/107
= /
= 0.8
/
= 0.1
/
=
=
These values are gathered from logged data of the hub motors. Those values
can also be gathered using a camera system to measure the speed at
corresponding positions. A low stopping distance has a good performance
result while a high MFDD gives a good performance result.
For category L1 (light-powered vehicle) the braking requirements are
summarised in table 2 [62].
Table 2 Braking requirement
Stopping distance MFDD
Front wheel(s) brkaing only S ≤ 0.1·V + 0.0111·V² ≥ 3.4 m/s²
Rear wheel(s) braking only S ≤ 0.1·V + 0.0143·V² ≥ 2.7 m/s²
Vehicles with CBS or split service
brake systems: for laden and
lightly loaded conditions
S ≤ 0.1·V + 0.0087·V² ≥ 4.4 m/s²
Vehicles with CBS – secondary
service brake systems
S ≤ 0.1·V + 0.0154·V² ≥ 2.5 m/s²
Since this vehicle is designed as a city delivery vehicle, the LMV will perform
above mentioned with different cargo amounts. The vehicle has a cargo
capacity of 200 kg. Tests will be conducted with cargo varying between 0 and
200 kg load.
The LMV’s primary breaking method is regenerative braking. The LMV is
also equipped with a front-wheel hydraulic brake. To evaluate the regenerative
braking, only the rear-wheel brake is applied. The braking energy charges the
battery and dissipates the remaining energy to heat using a brake resistor.
Braking with a high SOC will direct more energy to the braking resistor. These
tests are conducted on a high SOC and low SOC. Unfortunately, the efficiency
of storing regenerative braking reduces significantly at lower speeds [76].
32/107
4.3. µ-split braking test
The braking test is expanded upon for multitrack vehicles by braking on
different friction surfaces. This test is done at a velocity of 25 km/h. The
previously mentioned paper tests the braking on ice, but this is too difficult to
set up. The braking test will be performed on pavement on one side. The other
side is driven on loose gravel, steel sheets, polished concrete or low friction
tape around one tire to obtain a friction coefficient below 0.15 [77]. The
difference in friction will cause a different braking reaction on the left and
right wheels. Distortion after braking is measured to examine the stability of
the vehicle during an emergency brake as seen in figure 16.
Figure 16 µ-split brake.
The automotive industry uses ESC systems to improve control and reduce
collisions [78]. Cargo bikes are not equipped with ESC. City centres have
multiple areas with different road surfaces (i.e. tram rails, smooth tiles next to
the pavement, road markings, …) and µ-split brake can occur. A low distortion
indicates a stable vehicle with less likely to lose control.
33/107
4.4. Steady-state turn
A steady-state turn lets the driver drive around a circle with a radius of 4
meters as seen in figure 17. The driver drives as fast as possible around the
circle while remaining stable with all wheels on the ground.
Figure 17 Steady-state turn.
The maximum mean speed of at least one rotation is measured with its
corresponding lateral acceleration. A mobile phone is attached on the saddle
tube below the driver to measure the accelerations. A stable and controllable
vehicle reaches a high maximum vehicle. A low lateral acceleration means a
low sideways force applied on the driver and results in more comfort. A high
lateral acceleration causes a high outwards force which is pushing the driver
outwards and is highly uncomfortable. Depending on the centre of gravity and
the wheelbase, the vehicle remains stable during high lateral accelerations.
This manoeuvre tests the maximum velocity the vehicle can drive around the
circle.
34/107
4.5. Figure of eight
The figure of eight is an addition to the steady-state turn, where the control of
the vehicle is further examined. A figure of eight is driven around two pylons
standing 10 meters apart as seen in figure 18.
Figure 18 Figure of eight.
This measures the maximum mean speed of at least one complete eight with
the corresponding absolute mean acceleration. The speed will be lower than
the steady-state turn and the result will depend on the driving skills.
4.6. Prompt steering
Prompt steering evaluates the ability of sharp turns. Driving in city centres
often requires short turns due to narrow streets or avoiding obstacles. The
driver drives through two pylons at a constant velocity. The driver has to turn
as short as possible and the distance required to make a 90° turn is measured
as seen in figure 19.
Figure 19 Prompt steering.
These tests are conducted at 10, 15 and 20 km/h if the manoeuvre is deemed
safe. The steering is executed without braking. The minimal possible
35/107
distance should be measured at 2 meters after the turn, but due to the long
wheelbase of the LMV, the distance is measured when the vehicle is
perpendicular to the starting orientation. A manoeuvrable and controllable
vehicle will result in a short turning distance.
4.7. Double lane change
A double lane change and a slalom require quick changes and proper control
of the vehicle. The double lane change examines the highest possible speed
for the vehicle driving through the obstacle. The maximum mean velocity of
the course is measured. An adaptation was made in the manoeuvre to account
for wider vehicles as seen in figure 20.
Figure 20 Double lane change.
The consecutive bends review the controllability and stability. Since the
vehicle must drive between the pylons, extra width of a vehicle requires
extra sway to accomplish the manoeuvre making it more difficult. This test
examines if it’s possible to complete the course error-free at the desired
speed.
36/107
4.8. Slalom
The slalom is executed using fixed distances at 3 m for the short slalom and 4
m for the long slalom as seen in figure 21. The short slalom is done at 10 km/h
and the long slalom at 15 km/h.
Figure 21 Slalom.
The consecutive bends review the controllability and stability. Since the
vehicle must drive between the pylons, extra width of a vehicle requires extra
sway to accomplish the manoeuvre making it more difficult. This test
examines if it’s possible to complete the course error-free at the desired speed.
4.9. Summary
Above mentioned manoeuvres are summarised in table 3.
Table 3 Summary of manoeuvres
Manoeuvre
Performed speed
Tested aspect
Performance measure
Braking test 25 km/h
Control
Safety
Braking distance
µ-split braking
test 25 km/h
Control
Safety Distortion
Steady-state turn Max. speed Lateral stability
Max. speed
Lateral acceleration
Figure of eight Max. speed
Lateral stability
control
Max. speed
Lateral acceleration
Prompt steering 10, 15 and 20
km/h
Stability
Control
Manoeuvrability
Min. turning distance
37/107
Manoeuvre
Performed speed
Tested aspect
Performance measure
Double lane
change
Max. speed
Manoeuvrability
Max. speed
Slalom
10 km/h – 3m
cones
15 km/h – 4m
cones
Performance
Control
Manoeuvrability
Max. speed
Due to the nature of the vehicles, all manoeuvres are closed-loop manoeuvres
as the driver contributes a major part to the overall weight of the vehicle and
the centre of gravity of the vehicle. The driver's posture and grip change the
stability of the vehicle [79]. The manoeuvres which are said to be based on
tests from the automotive industry are similar to tests performed during a
motorcycle exam which evaluate the skills of the driver shown in figure 22.
These motorcycle manoeuvring test contains just as the bicycle tests: a slalom,
an emergency stop (brake testing), a figure of eight, avoidance (prompt
steering), cornering (steady-state circle). The driver requires certain driving
skills to pass these tests. Since drivers skills influences the results, the
manoeuvres are conducted multiple times and the best results are used as the
test are an indication of what is possible with the vehicle.
Figure 22 Motorcycle Manoeuvring circuit.
MOTORCYCLE
MANOEUVRING
Left Circuit
1 On and off the stand
2 Wheel the machine
3 Slalom
4 Figure of eight
5 30 kph/ 19 mph circuit
6 50 kph/ 32 mph avoidance
7 Controlled stop
8 U-turn
9 Slow ride
38/107
The prescribed test was performed with the LMV. These tests evaluate the
handling qualities and compare them with the benchmark vehicle.
All manoeuvres with the LMV are performed at a constant tire pressure of
6 bar for the rear wheels and 3.5 bar for the front wheel. These tests are
performed on the pavement. The tests are performed with the LMV with
and without load.
5. Results
A summary of the performed manoeuvres is described. The results of the LMV
are added to the results of the benchmark vehicles.
5.1. Braking test
The full application rear brake at 25 km/h is used to evaluate the regen braking
of the LMV. The current prototype has no place for mechanical rear brakes so
the vehicle only brakes using regen braking on the rear wheel hub motors. The
QuadRad has the best braking distances and is the only vehicle passing their
own acceptable braking distance as can be seen in table 4.
The maximum regulatory braking distance for a rear-wheel test at a speed of
25 km/h is 11.43 m [62]. This is acceptable for the City Shopping and the
Babboe Big. The regen braking has a proportional increase in braking distance
according to the weight progressing from 7.9 m to 12.92 m with no cargo and
200 kg cargo respectively. The braking distance is currently not sufficient with
a cargo of 160 kg or higher. The brake test only complies with the MFDD
guidelines only up to a load of 60 kg. This is due to the poor efficiency of
regenerative braking at lower speeds [76]. MFDD calculates the mean
deceleration during the braking. Adding a mechanical brake would greatly
improve the results. The battery’s SOC did not affect the braking distance.
Table 4 Full rear brake application 25 km/h
QuadRad City Shopping
Vm
(km/h)
LOAD
(m)
Sm
(m)
Skor
(m)
Acceptable
braking
distance
Vm
(km/h)
LOAD
(m)
Sm
(m)
Skor
(m)
Acceptable
braking
distance
25.3 180 6.7 6.54 √ 24.8 22 7.15
7.27
√
25.1 180 6.6 6.55 √ 25 22 7.5 7.6 √
25 180 6.6 6.6 √ 24.9 22 7.3 7.3 √
Babboe Big Babboe City
V
m
(km/h)
LOAD
(m)
S
m
(m)
S
kor
(m)
Acceptable
braking
V
m
(km/h)
LOAD
(m)
S
m
(m)
S
kor
(m)
Acceptable
braking
39/107
QuadRad City Shopping
distance
distance
24.7 80 7.30 7.48 √ / X
25.2 80 7.9 7.78 √ X
24.8 80 7.4 7.52 √ X
LMV
Vm
(km/h)
LOAD
(m)
Sm
(m)
Skor
(m)
MFDD
Acceptable
braking
distance
23.9 0 7.22 7.9 3.01 √
23.93 60 8.77 9.57 2.72 √
24.54 140 10.9 11.31
2.26 √
23.51 200 11.43
12.92
1.98 X
5.2. µ-split braking test
The µ-split brake of the LMV had distortions below 10 cm which indicates
that the LMV remains stable during braking. No data was available of the other
multitrack vehicles. The integrated ABS individually interrupts the
regenerative braking when a wheel starts slipping as can be seen in figure 23.
The wheel on the lower friction surface has more fluctuations in the wheel
velocity. This prevents the wheels from slipping and improves control during
braking. The µ-split braking test had on average a 10% longer braking distance
compared to the standard braking test.
Figure 23 µ-split brake hub motor speed.
124 124.5 125 125.5 126 126.5
Time [s]
-5
0
5
10
15
20
25
velocity [km/h]
Wheel on concrete
Wheel on gravel
40/107
5.3. Steady-state turn
The values of the steady-state turn can be seen in table 5. Babboe City was not
tested since this is a single-track vehicle. Babboe Big and City Shopping can
reach a velocity of 11 and 13 km/h and lateral acceleration of 2.14 and 3.02
m/s² respectively (figure 24 and 25). The QuadRad and the LMV reached a
velocity of 17.67 and 17.50 km/h. The lateral acceleration achieved was 4.5
m/s² and 0.89 m/s² (figure 26 and 27). Although the achieved velocity of the
QuadRad and the LMV is similar, the lateral acceleration of the LMV is much
lower, due to the ability to lean. The leaning shifts the centre of gravity
inwards, which is more comfortable for the driver. The lateral acceleration of
the cargo is 6.31 m/s² as seen in figure 26. This large lateral acceleration on
the cargo indicates the need for a low cargo area, which results in a low centre
of gravity and a stable vehicle. During the maximum speed, the LMV
remained stable and was still not tipping over. Driving with and without cargo
has no notable difference in maximum velocity. However, the extra weight
stabilises the rear module and improves driving comfort.
Table 5 Steady state-turn
Vehicle
V
m
(km/h)
Lateral
acceleration (m/s
2
)
QuadRad 17.67 4.5
City Shopping 13 3.02
Babboe Big 11 2.14
Babboe City / /
LMV 17.5 0.89
Figure 24 Lateral acceleration of City shopping during steady-state turn
[71].
41/107
Figure 25 Lateral acceleration of Babboe Big during steady-state turn
[71].
Figure 26 Lateral acceleration of QuadRad during steady-state turn [71].
Figure 27 Lateral acceleration of LMV driver and Cargo during steady-
state turn.
driver
cargo
42/107
5.4. Figure of eight
No results were available of the benchmark vehicles. The LMV has a mean
maximum velocity of 12.8 km/h. The respective lateral acceleration is 1 m/s².
Switching direction causes a lower maximum speed compared to the steady-
state turn. The lateral acceleration of the driver during the eight is higher as
shown in figure 28. This may be because the driver leans less due to the
changing turns. The maximum acceleration of the cargo is similar as during
the steady-state turn.
Figure 28 Lateral acceleration of LMV driver and cargo during figure of eight.
5.5. Prompt steering
The values of the Prompt steering can be seen in table 6. The Babboe City did
not participate in the steering input test because it is a single-track vehicle.
The extreme steering tests the behaviour of the multitrack vehicles. The
flexible frame allowed the LMV to lean during the turn preventing tipping and
keeping all wheels on the ground. Due to the long wheelbase, The LMV has a
bigger distance even though the turning clearance circle is smaller. The
QuadRad and the LMV also performed the test at 15 km/h with a distance of
1.9 m and 2 m respectively. The other vehicles were only carried out at 10
km/h for safety reasons. The LMV has similar results with and without cargo
but felt more stable with cargo. Driving without cargo has less traction and
causes some skidding of the rear wheels while turning.
driver
cargo
43/107
Table 6 Prompt steering
QuadRad City Shopping
V
m
(km/h)
A
min
(m)
A
min_kor
(m)
Tilting
V
m
(km/h)
A
min
(m)
A
min_kor
(m)
Tilting
-L 10.2 1.3 1.25 No 10 1.6 1.6 No
L 9.7 1.1 1.17 Delay 11 1.6 1.32 No
R 9.9 1.15 1.17 No 10 1.6 1.6 Yes
R 10.5 1.25 1.13 Delay 11 1.9 1.57 Yes
Babboe Big LMV
V
m
(km/h)
A
min
(m)
A
min_kor
(m)
Tilting
V
m
(km/h)
A
min
(m)
A
min_kor
(m)
Tilting
L 10 2.4 2.4 No 9.69 1.5 1.60 No
L 10 2.5 2.5 No 9.97 1.4 1.41 No
R 12 2.8 1.94 Yes 9.79 1.3 1.36 No
R 11 2.4 1.98 No 9.81 1.35 1.40 No
5.6. Double lane change
QuadRad can execute the double lane change error-free up to 25 km/h. The
single-track Babboe city can ride through the course at a maximum speed of
20 km/h. City shopping and Babboe Big can perform the course at 20 km/h,
but the tires already lose contact at 15 km/h. The LMV can perform the double
lane change error-free at a maximum of 20 km/h. The large width of the LMV
makes it considerably more difficult to drive through the course error-free as
can be seen in table 7.
Table 7 Double lane change
QuadRad
City
Shopping
Babboe
Big
Babboe
City
LMV
V
m
(km/h)
Course
passed
Course
passed
Course
passed
Course
passed
Course
passed
10 Yes Yes Yes Yes Yes
15 Yes Take off Take off Yes Yes
20 Yes Take off Take off Yes Yes
25 Yes / / / Yesa
a. This was carried out on the adapted double lane change
Due to the low clearance on the manoeuvre, the maximum possible speed
depends on the driver’s driving skills. The best results from the drivers differ
5 km/h. No significant differences were found using different loads. The
44/107
manoeuvres could be performed at the same velocity with and without cargo.
The drivers stated that the LMV was stable and all wheels remained on the
ground at all times. The tests with cargo felt even more stable compared to
those without. The cargo lowers the centre of gravity and improves the traction
of the rear wheels. At higher velocity, the driver is unable to drive through the
course without running over cones. The double lane change is also performed
on a modified course as explained on figure 20 so clearance is no issue. With
these conditions, the driver is able to go through the course with a maximum
velocity of 25 km/h. During this manoeuvre, the vehicle remains within a track
width of 2 meters. These passages carried out at maximum speed were error-
free on a course with 0.5 meters of extra clearance. In all tests, the LMV’s
wheels did not lose contact with the ground. This is thanks to the low centre
of gravity and the flexible connection between the driver module and the load
module.
5.7. Slalom
All vehicles except the LMV could perform both slaloms at prescribed
conditions. The LMV was able to perform the 4 m slalom at 8 km/h and the 3
m slalom was not executed without knocking over pylons. These results are
due to the high width of the vehicle. The high vehicle width requires stronger
manoeuvres compared to a narrower vehicle and the long wheelbase make the
3 m slalom too difficult.
6. Discussion
The various manoeuvres expose the strengths and weaknesses of the LMV’s
dynamic behaviour. The steady-state circle and the double lane change show
the LMV is a controllable and stable vehicle. The low cargo hold lowers the
centre of gravity when the LMV is loaded and improves stability. The flexible
connection between the driver and cargo lets the centre of gravity shift inwards
during turning. This ensures proper stability during extreme manoeuvres and
will prevent tipping over. During every manoeuvre, all wheels kept contact
with the road which was not the case with other multitrack vehicles. The
flexible connection also allowed for short turning without losing surface
contact as is noted in the turning clearance circle and the prompt steering.
However, the vehicle’s width creates extra difficulties during the manoeuvres.
The LMV has to take sharper turns compared to the other vehicles when
avoiding obstacles. These problems complicate driving in narrow city streets
or on bicycle lanes. In combination with the long wheelbase, this made the
short slalom not possible. The delta shape concept makes it difficult to
estimate whether the vehicle can pass without knocking over cones when there
is little room to manoeuvre.
45/107
All tests were performed on a flat asphalt surface since this is the majority of
the road the vehicles mainly drive on. However, the vehicle will be deployed
in city centres and will drive over cobblestones or through potholes. This test
does not consider the influence of bad roads and would be a great addition to
the setup where the accelerations are measured driving over different terrains.
Multitrack vehicles experience additional lateral shocks compared to single
track vehicles driving over an asymmetrical obstacle. These shocks are
extremely uncomfortable for the driver. The flexible connection on the LMV
absorbs these shocks for the driver. Furthermore, low-speed tests to control
stability are not part of the stability tests. These are not useful for the fixed
multi-track vehicles, but the single track and flexible multitrack vehicles can
evaluate the stability of this setup. Busy city streets will force the driver to
cycle at walking speed. Preferably this should be possible without putting feet
on the ground.
The LMV’s vehicle dynamics can be improved upon by changing certain
aspects of the vehicle design. Reducing the wheelbase will ensure an even
smaller clearance turning circle. This will facilitate the short turns done in the
manoeuvres and may improve the results. Furthermore reducing the width of
the vehicle would improve the clearance when driving through narrow
courses. The width at the level of the wheels cannot be reduced, but the cargo
hold can be narrowed towards the end. This ensures the vehicle swings out
less when cornering.
The braking tests are conducted using regenerative braking without the aid of
a mechanical brake. These results are not within the regulations when fully
loaded. The regenerative braking still brought the cargo bike to a controlled
stop whereas previous research stated brake testing with Babboe city was
stopped due to the vehicle being uncontrollable. To improve this, a different
hub motor with a higher maximum braking torque or adding a mechanical
brake is required.
The LMV has a decent result compared to the benchmark models. Although
the benchmark models represent the different types of cargo bikes, the delta
variant (city shopping) does not belong to a commercially usable category. A
better comparison can be done by testing additional cargo bikes with similar
characteristics. Cargo bikes such as Fulpra, Ono and Cargo Chariot. Those
examples have a flexible delta design similar to the LMV. However, these
models are hard to come by and currently very little represented in traffic
7. Conclusion
The manoeuvres prove the cargo bike concept has proper vehicle dynamics
independent of the cargo load. Apart from the braking, the results of the
manoeuvres were similar with or without a load. The drivers stated the vehicle
46/107
was even more stable with the load. This is thanks to the low cargo hold. The
concept disconnects the drivers module with the cargo module, providing
similar cycling feeling to that of a conventional bicycle. The concept vehicle
provides extra comfort compared to fixed multitrack vehicles. The hinge
allows leaning and significantly lowers the lateral force of the driver compared
to fixed frame multitrack vehicles.
The manoeuvres expose the weaknesses of the vehicle’s dynamic behaviour.
These weaknesses were used to determine various points of improvement.
The concept vehicle does not outperform current cargo bicycles in terms of
handling qualities but can be deployed for delivering higher weight and
volume parcels. The LMV handles less compared to the QuadRad since the
QuadRad has a smaller width and the manoeuvres are beneficial for narrower
vehicles.
47/107
Chapter 3: Pedalling comfort of a custom Pedal
Series Hybrid Drivetrain in a cargo e-tricycle
This chapter is based on the paper “Pedalling comfort of a custom Pedal
Series Hybrid Drivetrain in a cargo e-tricycle” published in the journal
“Engineering Science & Technology” [80].
1. Introduction
A new concept vehicle for last-mile delivery in city centres has been
developed [69]. A custom Pedal Series Hybrid Drivetrain (PSHD) was used
as the means of propulsion. This drivetrain powers the vehicle using
pedal-generated energy and battery-stored energy. The pedals drive a
generator mounted on the bottom bracket of the tricycle. Both rear wheels of
the tricycle include an electrical wheel motor. The battery is electrically
connected to the motors and the generator.
The lack of a conventional mechanical transmission, including a chain, belt,
or shaft, reduces the overall design constraints for the vehicle. Maintenance
interventions are expected to be reduced as well. However, a successful
market introduction of such a PSHD cargo tricycle also depends on its
acceptance by the couriers. A natural feeling of pedalling is a major
prerequisite. Therefore the generator resistance should vary during a
revolution of the pedals. A programmable resistance controller is developed.
The control loop uses the angular pedal velocity as a primary parameter and
is further finetuned using the propulsion values of the vehicle.
The technology, advantages, disadvantages, and examples of PSHD on the
one hand, and the state of the art in pedalling mechanics measurements on the
other hand, are described to illustrate the context and purpose of the study.
2. Pedal Series Hybrid Drivetrain
Traditional bicycles are powered only by the energy of the cyclist applied on
the pedals. It is transferred to the rear wheel by a chain transmission or
something similar. Only mechanical energy was used for the propulsion.
Conventional e-bikes have an electric motor to assist the cyclist. The
assistance motor is a mid-drive motor mounted at the bottom bracket or a hub
motor in the front- or rear wheel. E-bikes use the same cyclist power but are
also electrically assisted. The electrical assistance and human power work in
parallel. Therefore, the drivetrain in such an e-bike is called a Pedal Parallel
Hybrid Drivetrain. In a Pedal Series Hybrid Drivetrain, there is no mechanical
48/107
transmission between the pedals and the wheels. The only power transmission
is electrical. The pedals are connected to a generator that converts human
power to electrical energy [81]. This electric energy is stored in batteries or
capacitors or is used directly by one or more motors that propel the vehicle.
Most motors can be used, but BLDC hub motors integrated into a wheel are
the most common. This energy flow can be seen in figure 29.
Figure 29 Pedal Series Hybrid workings [64].
2.1. Advantages
The PSHD also has multiple advantages. The use of a PSHD is maintenance-
friendly [82]. The generator and the hub motors are both BLDC motors. These
types of motors are very low maintenance [83]. A conventional e-bike uses a
chain that requires a lot of maintenance and is one of the components most
vulnerable to wear. Furthermore, the majority of E-bikes are assisted using a
mid-drive. This configuration transfers the human pedal and drive power
through the chain. The combined power causes extra stress on the chain and
increases the chance of failure. An alternative that is less susceptible to wear
or failure is belt-drive, which is more expensive. The PSHD drivetrain reduces
the maintenance costs and is less likely to break down.
An electrical drivetrain has fewer design constraints compared to a traditional
chain transmission. The lack of a mechanical transmission creates more design
possibilities. Only a flexible electrical cable is required to connect the
generator with the propulsion. This “cycle by wire” connection requires no
alignment and doesn’t require room for a mechanical connection to the rear
wheels. Without a mechanical transmission, the wheel dropout can be smaller,
resulting in a narrower vehicle. The improvement of design constraints doesn’t
49/107
specifically benefit conventional bikes as they have evolved with a traditional
chain drivetrain. Furthermore, the drivetrain can operate with a virtual CVT
without using complex mechanical components. This allows the pedal
generator to be used constantly at the most comfortable cadence independent
of the vehicle’s speed, even while departing. E-bikes that use a rotation sensor
must depart on full human power before the assistance begins. These features
make the PSHD very ergonomic.
Recumbent bicycles have a long chain from the pedals to the rear wheel as a
drivetrain. They have a fixed seat and an adjustable bottom bracket to fit the
cyclist’s length. Adjusting the position of the bottom bracket also requires an
adjustment of the chain length, while an electrical drivetrain requires no extra
adjustment. This drivetrain can also be used in tandem. A conventional tandem
utilises two bottom brackets which are synchronized using a chain. As a result,
the cyclist must pedal at the same cadence. Replacing at least one pedal crank
with a pedal generator allows both cyclists to pedal at their comfortable
cadence.
PSHD makes pedalling stationary possible to recharge the battery, as
presented in the Lean Mean Machine [84]. This is only the case for multitrack
vehicles that remain upright while stationary.
2.2. Disadvantages
PSHD is an uncommon drivetrain. This doesn’t often occur since it has
multiple disadvantages. This type of drivetrain requires a generator and a
motor. These components can be more expensive compared with the required
parts on a conventional e-bike. However, if the e-bike is equipped with
internal hub gears or CVT gearing, which is implemented in the PSHD
drivetrain, the cost can be similar [82]. Hub motors have limited torque, and
the PSHD doesn’t receive additional torque from the pedals. European
regulations limit the nominal continuous power of e-bikes to 250 W [35]. The
motor power may be insufficient in uphill terrain for fully loaded cargo bikes
without the additional cyclist power that can provide high torque at low speed.
Furthermore, the pedal generator requires a custom control to experience a
natural cycling feeling. If the pedal generator doesn’t have a natural pedalling
feeling, the vehicle won’t be comfortable and won’t be used.
2.3. Examples of PSHD
There are only a few examples of vehicles using a PSHD, although the concept
was patented in 1975 [81]. The Mando Footloose is presented in 2013 and
uses a pedal generator and a rear-wheel BLDC motor. The lack of a chain
allows for a unique, easily foldable design, as seen in figure 30. The innovative
design even won the “2013 red dot design award” [85]. However, according
50/107
to reviews due to the limited torque, a noisy motor and a high price, the bicycle
wasn’t a success [86].
Figure 30 Mando footloose [85].
The Tortuga XL is a cargo bike using a PSHD and was presented in 2016 as
seen in figure 31. The cargo bike has a high cargo capacity. However, the
vehicle’s pedal generator has no variable resistance. The uncomfortable
pedalling feeling explains why the vehicle was not successful.
Figure 31 Tortuga XL [87].
A tadpole (two wheels in front) recumbent tricycle was built and tested in 2020
during the France Sun Trip [88-90]. On top of the pedal generator, PV solar
panels are used for power generation. Capacitors are used for energy storage
(figure 32a). A 1 kW hub motor is mounted in the rear wheel. The mechanical
decoupling of the pedals and wheels allows the cyclist to keep pedalling
without additional effort in mountainous terrain. It also allows the vehicle to
be pedalled stationary.
51/107
The recumbent tricycle of Bernard Cauquil also participated in the Sun Trip
(figure 32b). This delta tricycle uses a 1 kW hub motor in the front wheel [91,
92].
Figure 32 a) Vehicle of Edgar Tournon [89] b) Bernard Cauquil [92].
Another example uses 12 pedal generators that all work independently as seen
in figure 33. All cyclists can pedal in their comfortable cadence. This vehicle
is being tested in Le Teil, France [93].
Figure 33 Bicycle bus [93].
The above examples use custom-developed drivetrains. Mando developed the
Smart Personal Mobility (SPM) (figure 34a) and Schaeffler developed the free
drive (figure 34b) [94, 95].
a)
b)
52/107
Figure 34 a) Mando SPM [94] b) Schaeffler free drive [95].
2.4. Pedal Properties
To ride a bicycle, numerous actions are needed. Cyclists have to maintain
balance by steering and leaning whilst cycling. At the same time, they have to
apply force on the pedals for propulsion. These actions are automatically
performed after enough practice. Maintaining balance is aided by the self-
stabilizing properties of a bicycle. It even allows cyclists to ride without hands
on the handlebars. These self-stabilizing properties have been researched
thoroughly [70, 75, 96-101]. It depends, among other things, on the caster of
the fork and the gyroscopic effects of the wheels. The complete geometry and
mass of the vehicle have further influence. The effect varies with the bicycle’s
velocity. The stability of a bicycle can be simulated using the Carvallo-
Whipple model [70].
Although the cyclists interact with their seats, handlebars, and pedals [68, 102-
104], the majority of the cyclist/bicycle interaction research is focused on the
global forces exerted on the pedals since this creates the propulsion [29-34].
Research reveals the force applied on the pedals varies according to the
orientation of the pedal cranks. These variations are due to the biomechanical
structure of the human leg [105, 106]. A leg can generate a high downward
force but can only to create a limited horizontal force [107]. Since the feet are
not fixed on a standard pedal, no upward force can be applied to the pedals.
The feasible forces of the leg can be seen in figure 35.
b)
a)
53/107
Figure 35 Set of feasible forces at different crank angles [105].
This corresponds to pedal measurements [68, 107-111]. The highest force is
applied during the downward pedal movement [105]. During the horizontal
and upward pedal movement, the cyclist can not apply a useful force on a
traditional bicycle, as can be seen in figure 36.
Figure 36 Right pedal forces and torque profiles (a) Fz (b) Fy (c) Fx (d)
torque with 0° being the upward pedal position [111].
Since the pedals alternate, the upward movement of one pedal is covered by
the downward movement of the other pedal. Around the horizontal pedal
54/107
movement, corresponding to a vertical pedal crank orientation, little to no
force is applied by both legs. These vertical pedal crank orientations are
therefore called dead centres. They are however easily overcome due to the
bicycle’s inertia. The force can be converted to torque and the dead centres
are also clearly visible in figure 37.
Figure 37 Crank torque measurement by the pedals and SRM torque
analysis [110].
The variation in the crank torque shows two peaks in one rotation each created
by one pedal. The two minima represent the dead centres. Although the
bicycle’s inertia overcomes the dead centres, this can still cause a small
momentarily angular decrease of the pedal cranks. An angular velocity
increase can occur during the torque peaks.
The majority of research assumes that the difference between the
Instantaneous Angular Velocity (IAV) and the average pedalling cadence is
negligible and is therefore rarely mentioned. Force and torque measurements
don’t require an angular velocity but should know the crank orientation. These
orientations are often interpolated using the cadence. Power meters also
require an angular velocity. Current power meters utilize the average cadence,
which can cause a miscalculation. Favero Electronics developed an IAV
Power meter [112] to improve the accuracy (figure 38a). Their research
noticed an IAV variation while cycling on a flat road between 5% and 10%
(figure 38b). This variation can cause a power miscalculation as high as 1.6%.
The difference can be significantly greater under certain conditions such as on
a stationary trainer due to the lower inertia or when using an oval chain ring.
55/107
Figure 38 a) Force and angular velocity of IAV power meter [112] b)
angular velocity and tangential force of a single pedal revolution of 5
different cyclists [113].
3. Research objectives
Although the PSHD concept has been known since 1975 [81], no significant
penetration of this type of drivetrain has been observed in the market. The
main objective of this research is to prove the suitability of a PSHD as an
alternative drivetrain for bicycles. To reach this objective, it should be shown
that the three primary disadvantages of a PSHD are irrelevant or can be
overcome:
• a PSHD should have a natural or comfortable pedalling feeling.
• the generator can serve as an electrical gearbox, reducing mechanical
components such as a derailleur, internal hub gearbox, or something
similar.
• the cyclist should experience unfavourable driving conditions such
as driving fully loaded, driving uphill, driving into a headwind, …
and have the choice to add more power under such conditions.
a)
b)
56/107
4. Tricycle description
The vehicle used in this research is a tilting delta tricycle (figure 39) developed
after extensive market research [9]. This cargo bike can carry euro pallet sized
loads with a maximum payload capacity of 200 kg. The vehicle is designed to
be stable and manoeuvrable, independent of the payload. The tricycle consists
of a tilting front driver module that allows the driver to tilt while manoeuvring,
resulting in a comfortable driving feeling [69].
Figure 39 Delta tricycle cargo vehicle.
4.1. Hardware
The delta tricycle utilizes a PSHD. Both rear wheels were equipped with a
2000 W BLDC hub motor. They have hall sensors for incremental positioning
and speed measurements. The generator mounted on the pedal bracket is a
250 W mid-drive motor modified to allow the pedals to drive the BLDC
motor. It also uses hall sensors for speed measurements. Custom control
hardware and software are used for the generator and both hub motors. An
ESP32 data link can connect to a smartphone that serves as a dashboard for a
real-time view of velocity, energy consumption, energy storage, settings, …
4.2. Control loop
Driving the vehicle requires two control loops. One control loop operates the
generator, and one control loop operates the hub motors in the wheels. The
pedal generator control loop regulates the power generation to achieve a
natural pedalling feeling.
In a mechanical drivetrain, a balance is present between the applied pedal
power and the acceleration of the vehicle. Increasing the pedal power results
in an increase in the vehicle’s velocity. Lowering the pedal power decreased
57/107
the velocity. On the other hand, changes in the environment like an increase
in slope require an increase in pedal power to maintain a constant velocity.
The electrical drivetrain simulates this effect through data exchange between
the generator and the hub motors. Vehicle speed is derived from the hall
sensors in the hub motors (wheel_speed). The average current (wheel_current)
of the hub motors indicates the used propulsion.
Wheel_speed is scaled to the rotational speed of the pedals using a virtual
gearing (kph2rps). The used hub motor current is scaled using the same virtual
gearing. The wheel current is further scaled according to the desired
Assistance Level. This results in a scaled amplification factor for the control
loop.
The rotational speed difference between the geared wheels and the pedals is
proportionally scaled to a current value. A constant counter-torque is added to
maintain some small resistance during the dead centre. The scaled wheel
current is also added with this and results in a current setpoint. The change in
the current setpoint is limited to reducing the jerk in the pedal generator. This
control loop can be seen in figure 40. The control loop results in a natural
pedalling experience and gives proportional feedback according to the wheel
power. Furthermore, this control system serves as a continuous variable
transmission. Gearing changes automatically to provide the driver with a
comfortable pedalling cadence.
Figure 40 Pedal generator control loop.
The control loop for the propulsion is done individually for each hub motor.
The control loop receives the pedal current setpoints. The setpoint is scaled
using the gearing and the Assistance Level. An integrator of the speed
difference between the speed of the hub motor and the maximum velocity is
added as a velocity limiter. This results in a current setpoint for the hub motor.
The control loop can be seen in figure 41.
58/107
Figure 41 Wheel control loop.
The pedal generator should create a natural feeling without feedback from the
pedal crank orientations. The rotational difference between the pedals and the
hub motors is used to simulate the inertia present in a traditional drivetrain.
However, using the rotational difference requires a noticeable difference with
limited gain or the pedal generator would have an uncontrollable oscillating
resistance. This can cause a bigger difference in the rotational pedal speed.
5. Methods
The PSHD must create the natural feeling of riding a mechanical drive train.
To evaluate and improve this natural feeling, test subjects used the test tricycle
and commented on their experiences. The test subjects were randomly
selected. While driving the vehicle, data from the pedal generator were
measured and stored. These data are compared to data gathered by driving a
conventional e-bike.
5.1. Pedal generator
The relevant data to evaluate the pedal generator is collected whilst cycling on
bicycle lanes. This test collects the motor position, the motor speed and the
motor current according to the time. The data is processed to organise the
motor speed and the motor current according to the pedal orientation. The
values are interpolated to obtain an equal spacing of 1 degree. As both pedal
cranks are attached to the same motor shaft, the generated motor current
accounts for the force applied on both pedals. This generated current is
proportional to the applied tangential force. The motor speed is scaled to the
internal gearbox of the pedal generator to obtain the pedal cadence. This pedal
cadence is further analysed. The instantaneous cadence is divided by the mean
cadence of the same revolution, as seen in equation 3.
=
(3)
59/107
With:
=
=
= %
The result displays the angular velocity in percentage and shows the variations
in the cadence according to the crank orientation.
5.2. Measurement bicycle
The pedal generator is compared using the pedal values of a measurement
bicycle. The measurement bicycle measures the complete cyclist/bicycle
interaction. The bicycle has a 6-dimensional load cell incorporated in the seat
post [102] and the handlebar stem [103]. There is also a 3-dimensional load
cell in each pedal [68]. The orientation of the pedals and the crank are
measured, enabling the calculation of the tangential crank forces for each
pedal. The tangential forces of the pedals are added up to the total tangential
force exerted on the mechanical drive. The data from the measurement bicycle
were obtained from Dieltiens et al. [114]. Dieltiens et al. used the data to
analyse differences in cycling with and without assistance. The data of a pedal-
assisted test is collected and organised according to the pedal orientation. The
values are interpolated to acquire a one-degree interval from 0° to 360°. The
crank positions according to time are used to calculate the angular velocity of
the pedal crank. Using equation 3, the variations in the angular velocity
according to the crank orientation are obtained. The measurement bicycle was
set up stationary on resistance rollers. Resistance rollers are used to stationary
simulate a resistance similar to natural cycling.
5.3. Comparison method
The generated pedal current is compared with the total tangential force of the
measurement bicycle. They don’t have the same unit and require a scale for
comparison. The data processed according to the pedal orientation allows for
the same size comparison. Furthermore, the variation in the angular velocity
is compared. The combined analysis determines the differences between the
pedal series hybrid and a pedal parallel hybrid. A Student's T-Test is
performed with the scaled current data of the pedal generator and the total
tangential pedal force of the measurement bicycle. This test evaluates the
significance between the resistance and the tangential force.
6. Results
All test subjects were impressed by the performance and natural feeling of the
test tricycle. They stated that pedalling the tricycle feels natural once departed.
60/107
They used the level of assistance just like with a conventional e-bike. They
experienced uphill and full-load driving conditions. Departing from zero
speed took a couple of tries to get used to it. They obviously had to get used
to the unexpected behaviour: the pedals can be rotated even when the tricycle
does not move yet. After some attempts, this difference from a traditional
drivetrain is experienced as very favourable: almost no force is required to
depart. Furthermore, riding at a preferred rotational pedal speed is just as
natural as riding on a conventional bike. About half the test subjects mentioned
a difference between driving the generator and a conventional e-bike without
being able to describe the difference in detail. Some of them mentioned
“pushing through the resistance”. This difference is contributed to the higher
gain in angular velocity around the foremost position of the crank.
The measurements selected in figures 42 and 43 have a cadence between 50-
70 rpm. This was done at a constant velocity between 15-25 kph. The selected
data from [108] originate from a single level of assistance and the
measurement bicycle did not shift gears.
The pedal generator data consist of 1703 full pedal rotations and the
measurement bicycle data consists of 124 full pedal rotations. The graph
shows the mean ± 2 times the standard deviation of the pedal generator torque
in grey and the measurement bicycle scaled tangential force in red according
to the pedal orientation. 0 degree is the vertical orientation of the pedal cranks
with the right pedal upwards. The force/resistance of the first peak was applied
by the right leg and the second by the left leg.
Figure 42 Force/current according to orientation and T-Test results.
Mean LMV ±2 × std
Mean instrumented bike ±2 × std
deviance
above 5% significance
61/107
Figure 43 Graph of angular velocity variations and T-Test results.
7. Discussion
The tangential force of the measurement bicycle clearly shows the same
behaviour compared with the pedal generator. The T-test proves the majority
of the first half has 95% confidence of significance. The pedal generator has
a smoother curve compared to a traditional bicycle. This is due to the control
loop that controls the resistance. The standard deviation of the pedal generator
looks proportional according to the resistance. The dead centres have a low
deviation as the control loop receives little input of the rotational difference
between the pedal generator and the main input is the constant counter torque.
The pedal generator has a rotational speed IAV variation of ± 10%. This is a
significant difference compared with a traditional drivetrain. The
measurement bicycle has a lower variance of ± 4%. Resistance rollers have
lower inertia than road cycling due to the moving mass of the bicycle and
cyclist, which is not taken into account in the measurement of bicycle setup.
This may have caused a higher variation in the angular velocity of the pedal
crank compared to cycling on the road. Testing on the roads would have
caused a greater difference. The flexible transmission of the pedal generator
and the vehicle wheel results in a higher variation in the angular pedal velocity.
The flexible transmission also causes a higher deviation in the variance. This
variation isn’t extreme as Favero noticed variations ranging from 2% to 10%
[113].
The IAV variation of the angular velocity corresponds to the generated
current. During the dead centres, the angular velocity is around -10% of the
mean cadence and the generated current is also at its lowest. At the orientations
where the pedal generator generates peak currents, the angular velocity is
around +10% of the mean cadence. The rotational difference between the
pedal generator and the hub motors simulates the inertia as observed during
the resistance. However, due to the set proportional gain and the electrical
Mean LMV ±2
×
std
Mean instrumented bike ±2 × std
deviance
above 5% significance
IAV [%]
62/107
transmission, wanting to push more on the pedal generator causes a
momentary increase in the rotational speed. This effect is only slightly present
on a conventional drivetrain. An increase in force has little effect on the
rotational speed due to the high inertia of the vehicle and the stiff mechanical
drivetrain.
There is an asymmetry present in the measurement of bicycle data that cannot
be explained properly. The peaks of the left leg and the right leg have an equal
maximum force, however, the peaks are 200 degrees apart. This should be
180° ± 1° [110]. Furthermore, the applied force of the left leg doesn’t have the
same sinusoidal shape as the right leg. The same asymmetry is present in the
IAV variation of the instrumented bicycle. The peak of the right leg occurs
around the same orientation while the peak created by the left leg is 20 degrees
after the pedal generator. This can be due to the difference in the left and right
legs of the cyclist.
8. Conclusion
The three main objectives of this research have been reached:
• Test subjects experience a natural pedalling feeling when riding the
test tricycle. This is confirmed by comparing pedalling data from this
tricycle with data from a pedal parallel drivetrain in a conventional
mid-motor e-bike. The mechanical resistance experienced by cyclists
when pushing their pedals is similar to the PSHD and the
conventional e-bike. Some test subjects reported an undefined
difference between the tricycle and a conventional e-bike. This
difference is attributed to a higher variance of the Instantaneous
Angular Velocity in the PSHD, caused by the choice not to use a
crank orientation measurement.
• The PSHD serves as an electrical gearbox by creating a variable
speed ratio between the pedal speed and the wheel speed. This has
the benefit of maintaining an optimal pedalling cadence.
• The actual global driving conditions are used in the control loop for
the resistance of the generator and influence the level of electrical
assistance based on the electrical current in the hub motors.
While the pedal generator offers a natural feeling, there remains room for
improvement to minimize IAV variation and achieve an even more natural
feedback and feeling simulating a rigid transmission.
63/107
Chapter 4: Effects of a Torsion Spring Used in a
Flexible Delta Tricycle
This chapter is based on the paper “Effects of a Torsion Spring Used in a
Flexible Delta Tricycle” published in the journal “applied mechanics” [115].
1. Introduction
In the transition towards sustainability, there is a shift in last-mile delivery
towards cargo bikes. Cargo bikes are cheaper, faster and more efficient for
last-mile delivery in city centres compared to conventional vans [13]. This
shift is visible with major courier companies like DHL, UPS and FedEx testing
cargo bikes in major cities like London, Berlin and Amsterdam with positive
results [53-55]. However, current cargo bikes still have a limited cargo size
and area which requires more trips [50]. Furthermore, cargo bikes are ideal for
densely populated areas when logistic hubs are close to the city centre [116].
Improving the cargo area and using city containers can reduce and facilitate
loading the cargo bike and eliminate the need for a transportation hub. Current
cargo bikes are primarily developed for private use. Novel cargo bikes are
developed for last-mile delivery and are able to transport city containers and
a maximum payload of 200 kg. These cargo bikes are already replacing
courier vans and have the potential to deliver 85% of all last-mile deliveries
[117]. A cargo bike was developed to improve the cargo capacity and handling
qualities [9]. This concluded in a tilting delta tricycle. The tricycle consist of
a tilting driver module and a level cargo module. Tilting allows for a
manoeuvrable vehicle. However the driver module can tip over. Furthermore,
the tilting design causes low speed instability just as on a conventional bicycle.
Different methods are reviewed to prevent tipping over and improve the low-
speed stability. A torsion system implements a counter-torque that prevents
tipping over and improves the low-speed stability. The influence while turning
is reviewed and calculated for and an optimal counter-torque is considered.
2. Concept
Cargo bikes require high comfort standards. The courier has to drive the bike
the whole day so proper stability and controllability are required. A heavy
payload increases the need for a stable vehicle which prevents rolling over
during manoeuvres and the desire for a low cargo hold. A wide wheelbase of
multitrack vehicles provides extra stability. Furthermore, cycling multitrack
vehicles on bad road conditions worsens the comfort due to lateral shocks
caused by cycling through potholes with one side of the vehicle. Without back
64/107
support as present in a car seat, the driver must be able to tilt during
manoeuvres to have a comfortable driving feeling. A proper suspension can
absorb the lateral accelerations due to bad road conditions but is a complex
and expensive component of the vehicle.
Creating a hinge between the cargo module and the driver module allows the
driver to lean in manoeuvres. This improves the driver’s comfort and allows
the driver to lean while manoeuvring similar to a traditional bike. An extra
advantage over long johns is the cargo load has a limited influence on the
handling qualities of the vehicle.
The requirements of the vehicle resulted in a flexible delta tricycle as seen in
figure 44 [9]. The concept vehicle consists of a cargo module and a driver
module that are connected with a hinge allowing the driver module to tilt. The
cargo module has two rear wheels each equipped with a hub motor for
propulsion. The cargo area is placed between the wheels at a minimal ground
clearance. The driver module is equipped with a pedal generator. The pedal
generator works as a generator and has a virtual gearing with the hub motors.
This design allows a wide and low cargo area with the ability to place a high
city container without obstructing the driver’s view. The flexible frame allows
the driver to lean during manoeuvres as with a traditional bicycle. The concept
has proven to have proper handling qualities [69].
Figure 44 Last-mile vehicle concept [69].
The extra degree of freedom allows the driver to shift the centre of gravity
inwards. It allows the vehicle to perform stable manoeuvres at high speed
without the risk of toppling over which is present with fixed frames tricycles
[118].
The handling qualities of the concept vehicle were tested through an extensive
set of manoeuvres and resulted in proper handling qualities [69]. The flexible
frame makes the vehicle extra nimble. However, the manoeuvres used were
tailored to multitrack vehicles without taking tilting vehicles into account.
65/107
These tests excluded stability tests at low speed. Traditional multitrack
vehicles are tri- or quad cycles with a fixed frame. Due to the wide wheelbase,
these vehicles will remain upright while manoeuvring at low velocities and
while stationary. A flexible delta tricycle has similar self-stabilizing properties
as a traditional bicycle. These self-stabilizing properties only have a minimal
effect at low speeds resulting in more effort for the driver to remain stable at
low velocity [70].
When the driver module is placed upright it can remain vertically whilst
stationary. However, this is at an unstable equilibrium. A small force can shift
the centre of gravity of the driver module outside of the wheelbase and result
in the driver module falling over. This is additionally caused by the rotation
of the steer. Rotating the steer moves the wheelbase even further from the
centre of gravity. Multiple methods to improve this concept in low-speed
stability and stationary stability are examined and weighed against each other
to obtain a solid result.
3. Tilting/Tipping over
DIN 79010 is currently the only cargo bike safety standard. This is however a
German standard and is currently not mandatory. There is no European
standard for cargo bikes, however, the future European standard will most
likely be based on the DIN standard as this has happened before for a similar
subject [119]. The standard described multiple aspects such as braking,
dynamic stability and parking stability. The parking stability describes the
testing procedure to evaluate the parking capability. The standard requires the
vehicle to be able to be parked at inclinations of 8% (4.6°) sideways and
longitudinally uphill and downhill of 10% (5.7°), as seen in figure 45. This
standard however does not state what is defined as fall/roll over for a flexible
multitrack frame. The driver module could completely fall over while the
cargo module remains upright. Furthermore, the driver module could tilt at a
certain angle, but no restriction is stated on what is allowed.
Figure 45 Lateral inclination, downhill inclination and uphill inclination.
66/107
3.1. Bicycle stand
Similar to a traditional bike, the driver module can be equipped with a side
stand that acts as extra support to prevent the module from tipping over at one
side. This is used in the “Cube concept dynamic cargo”, calderas one from
Sblocs, “electric Cargo Trike concept” from James Dyson Foundation and
“butchers and bicycles”. This is an obvious solution and can be equipped using
standard side stands used for bicycles. Other variants can be used to support
the cargo area while handling cargo.
The traditional side stand is placed behind the pedals or at the rear wheel and
when closed, it extends backwards. Mounting a traditional stand on the
concept driver module would collide with the wide cargo module.
Lengthening the wheelbase would create extra room but reduces the vehicle’s
manoeuvrability. Furthermore, using the stand can cause difficulties. The side
stand only prevents the vehicle from falling to one side. Just like with
traditional bicycles, when placed on a hill or uneven terrain, the driver module
could tip over. Moving or rotating the bicycle to park it stably is much easier
with a bicycle than doing the same with the concept vehicle due to the extra
weight and the bigger clearance circle.
This method only aids the stability of the bicycle while stationary. The low-
speed handling qualities are not affected by a bicycle stand. Due to the
reorganization of the city centre, cargo bikes can cycle through busy
shopping/walking streets and occasionally have to cycle slowly. Improving
the low-speed stability would be beneficial. Furthermore, a stand requires
opening and closing manually every time the driver steps off, which is a
substantial amount if the vehicle is used for last-mile delivery.
3.2. Locking mechanism
To keep the driver module upright, the vehicle can be equipped with a locking
mechanism. A locking mechanism can lock the tilting mechanism preventing
the vehicle to tilt. This is commonly used in flexible tadpole tricycles like
“Butchers and bikes”, “pro cargo CT1”, Chike, and “HNF- Nicolai CD1”.
Tadpole tricycles do not have a separate driver and cargo module. The cargo
area tilts with the cyclist just like a single-track vehicle. Poorly distributed
loads would cause unwanted tilting or falling off the vehicle. The locking
mechanism prevents this. Locking the tilting system offers the advantages of
a fixed frame multitrack vehicle. The wide fixed wheelbase ensures a stable
vehicle at low velocity and when stationary. This also allows the vehicle to
remain upright when unmanned and eases mounting and dismounting. The
driver still has to manually lock and unlock the tilting mechanism.
Furthermore, the vehicle is less nimble when the tilting mechanism is locked.
Cycling with or without the locking mechanism engaged is like cycling two
67/107
different vehicles. A locked tilting vehicle has the same handling qualities as
a fixed frame tricycle. The cyclist can forget if the locking mechanism is
engaged. This could cause the cyclist to take a corner too quickly and end up
in the vehicle tipping over.
An electronic locking mechanism can provide automatic locking. This
engages the mechanism automatically depending on the velocity. This way the
vehicle will always drive optimally. At low velocity and whilst stationary the
vehicle is locked. This ensures a stable slow drive and keeps the vehicle
upright when stopping and stepping off. At higher velocity, the vehicle can tilt
ensuring proper manoeuvring capabilities.
To the author’s knowledge, this locking mechanism has not been implemented
yet in delta tricycles. This is likely due to the lack of suspension in the cargo
module of the vehicle. While the fixed frame improves the stability at low
speed, the vehicle experiences uncomfortable lateral shocks when driving on
bad road conditions or over speed bumps due to the limited suspension similar
to fixed frame delta tricycles.
3.3. Angle stop
The use of a stop limiting the tilting angle is a way to prevent the driver module
from completely tipping over. This is automatically incorporated in the
majority of the tilting tricycles but varies in the angle limit. The angle limit
requires a high enough range, since limiting the tilting angle reduces the
manoeuvrability of the vehicle. This was tested by changing the tilting angle
of the concept vehicle. For example, changing the maximum tilting angle from
30° to 25° reduces the maximum possible speed when driving a circle of a
radius of 4 m from 17 km/h to 14 km/h. During evasion manoeuvres, the test
subject was not able to perform an evasive manoeuvre with the reduced angle
limit at the maximum velocity which was possible with an angle limit of 30°.
The test subject also stated the angle limit gives a sudden shock when reaching
the angle limit at 25°. This did not occur during the test using the 30° angle
limit. The shock of the angle limit can be reduced by adding a damper.
The results were unexpected since the majority of delta tricycles have a tilting
limit of 18–20° [120]. The tests concluded a 30° angle limit is required. This
is however an extreme position to pull/push the driver module upright before
getting on the vehicle. Tilting this far would also cause the steer to extend 40
cm outside the width of the vehicle. This can cause the steer to fall onto a
nearby vehicle.
68/107
3.4. Torsion system
A different solution is equipping the hinge with a torsion system. This torsion
system creates a torque resistance depending on the tilting angle. The
resistance will push the vehicle upright. The vehicle is still able to tilt and aids
in the stability of the vehicle at low velocity. This system is present in tadpole
tricycles as “the cargo e-bike concept ” of Volkswagen [121].
The torsion system in the concept vehicle is made up of a hollow square tube
with a smaller square tube placed inside and rotated 45°. The outer tube is
fixed with the cargo module while the inner tube rotates together with the
driver module. The free space is filled up with an elastomer as can be seen in
figure 46. This system is similar to a torsion axle used in trailers.
Figure 46 Torsion axle and corresponding spring characteristic.
A torsion system is beneficial for multiple reasons. The torque ensures that the
driver module remains upright and prevents it from tipping over. This
eliminates the need for a parking stand and eases mounting and demounting
while delivering parcels. The torsion system also acts as a soft angle limit as
the torsion increases according to the tilting. This will also improve the
stability of the vehicle at a lower velocity. The counter torque present during
tilting will push the driver module to the centre. This counter-torque will be
limited so it will not prevent the driver from falling over but will facilitate for
the driver to remain stable. The torsion system has an additional advantage.
The hysteresis present in this system acts as a damper reducing possible
vibrations that can occur while driving. The counter torque requires a certain
torque/tilt curve to provide proper aid in the handling qualities of the vehicle.
If the counter-torque is too low, the vehicle will simply fall over when not
supported and driving slowly requires just as much concentration as without
the torsion system. When the counter-torque is too high, the driver is only able
69/107
to tilt too little during manoeuvres. This causes the driver to endure extra
lateral acceleration similar to a fixed frame. The torque/torsion curve can vary
depending on the properties and size of the elastomers.
The concept vehicle performed the manoeuvres equipped with the previously
mentioned torsion system. The torsion spring does not influence the comfort
of the driver during high-speed manoeuvres [69].
Testing the handling qualities shows that manoeuvres such as a double lane
change can be performed at maximum velocity. During this manoeuvre, the
cyclist only experiences a lateral acceleration of 2 m/s2. This is a comfortable
lateral acceleration in a cycling seat position [15].
4. Used system
The vehicle concept incorporates multiple of the previously mentioned
systems as seen in figure 47. The concept vehicle incorporates a stand, torsion
system and angle limit in the design.
Figure 47 Tilting concept vehicle.
The concept vehicle has a tailboard on the back of the cargo module. The
tailboard can rotate downwards and support the cargo module. This prevents
the vehicle from tipping and raising the front wheel when loading and
unloading a container. The incorporation of a torque system and angle limit in
the hinge ensures a simple implementation in the concept vehicle. This system
is chosen because it will prevent the driver module from falling over without
the need to enable a locking mechanism. Furthermore, the torsion system can
also aid in low-speed stability whilst having little to no influence on the
manoeuvrability of the vehicle.
Tailboard
stand
Torsion system
Angle limit
70/107
4.1. Effects of torsion system on stationary stability
The concept vehicle is equipped with a torsion system that creates counter-
torque when tilting. This provides an incorporated soft angle stop which limits
the tilting and in extreme cases, an angle limit of 30° is used. A flexible torsion
system was implemented in the concept vehicle as high-speed
manoeuvrability has priority over low-speed stability. To acquire its
torque/angle curve a torque is applied in an upward and downward motion and
the angular deflection is measured. The graph, as seen in figure 48, is obtained.
A third-degree polynomic trendline is fitted on the upper side for further
calculations.
Figure 48 Measured torque/angle curve with corresponding upper
trendline t=0.0013·a³+0.016·a²+0.0851·a+2.3532. with t the torsion
(Nm) and a the torsion angle (°).
The elastomer torsion system has a typical hysteresis which provides a
dampening effect. This reduces vibrations while cycling. The currently
equipped torsion system has a zone between −5° and +5° where the torsion
system creates almost no counter-torque. This small counter-torque is still
enough to keep the driver module upright without a driver. However, when
disturbed or while handling cargo, the driver module tilts between 14° and
17°.
This is measured after creating disturbances by loading and unloading the
cargo hold. When the driver module is pushed to the angle limit, the driver
71/107
module tilts back to a maximum of 22° as shown in figure 49. Tilting 22°
accounts to the steer extending about 30 cm outside the width of the vehicle.
Figure 49 Tilting driver vehicle while parked.
The torque the driver module applies on the hinge is calculated using equation
4.
= ··sin()· (4)
With:
=
=
= 9.81 /²
=
= °
The centre of gravity distance is measured perpendicular to the hinge’s axis.
While parked, the driver module is unmanned and has a mass of 15 kg and a
centre of gravity distance of 0.3 m perpendicular to the hinge’s axis. While
cycling the driver module is manned and has a mass of 90 kg and as centre of
gravity distance of 0.8 m.
72/107
Equation 4 is however a simplification since there are multiple unknown
variables. The steer can rotate which moves the point of support and creates
an asymmetry in the centre of gravity. This has a bigger effect when the driver
module is upright. When the driver module is tilted, the change is not that
great. The steer is even pushed straight thanks to the geometry of the vehicle.
Due to the uncertainty of the steer rotation and the small effect it has, this is
neglected.
Furthermore, the weight distribution of the cargo module can create an
additional force on the driver module. When the centre of gravity of the cargo
module is inclining forward, the front wheel receives additional weight. This
can create an additional torque in the hinge. In this concept, this has a limited
effect due to the axis of the hinge intersecting with the contact point of the
front wheel when positioned upwards.
The calculated moment created by tilting on a level surface intersects with the
counter-torque at 13.25° as shown in figure 50. When the vehicle is at a
diagonal inclination of 8%, the driver module hinges between 18.5° and 22.5°,
resulting in a vertical inclination of 23.1° and 27.1°.
Figure 50 Stationary tilting curve on a level and inclined surface.
These calculations correspond to the measured values. Although the driver
module stops tilting before the angle limit is reached, the driver module still
73/107
tilts a decent amount which creates a tedious action when getting on the bike.
Furthermore, the steer can extend 30 cm outside the vehicle’s width and fall
onto close by vehicles. Increasing the counter-torque in the hinge will reduce
this.
The torsion system has an influence while cycling. During manoeuvres, the
driver tilts while cornering. Cornering causes the driver to lean inwards to
compensate for the centripetal force. Cornering the vehicle at a certain radius
at a certain velocity causes a centrifugal force as calculated in equation 5.
=×
(5)
With:
=
=
= /
=
This centripetal force acts on the driver module and the driver pushes it
outwards of the turn. Tilting inwards counteracts this thanks to gravity as seen
in equation 4. This creates a total torque in the hinge stated in equation 6.
=
×cos()×– ××sin()× (6)
With:
T= in
=
=
Without a torsion system is 0 Nm. When calculating it with the torsion
system, is substituted by the torque/angle curve. This results in a decrease
in tilting due to the counter torque. equation 6 is compared with measurements
performed on the concept vehicle. Using measurements of different
manoeuvres corresponding tilting angles are gathered according to the turning
radius and the velocity. Multiple manoeuvres are performed with the vehicle
and turns are sorted according to turning radius and velocity. Figure 51 shows
the measured tilting angle according to the turning radius. The colour is plotted
according to velocity.
74/107
Figure 51 Tilting angle according to turning radius and velocity.
Calculations are performed at 5, 10, 15, 20, and 25 km/h.
The measured and calculated tilting angles have a mean difference of 2°.
Those differences can be explained due to simplifications in the equation. The
driver who accounts for the majority of the driver module’s mass is considered
rigid while this is not the case. During manoeuvres, the driver can rotate its
torso, changing the centre of gravity. Taking a turn is never performed at a
constant turning radius and velocity. Selected measurements are steady turns
with a limited variation. The mean values during the turn are used for the
calculations. The calculations are a simplification where the movement of the
wheelbase due to steering and tilting is not taken into account as well as a
slightly smaller turning radius of the driver module compared with the cargo
module which is used to calculate the turning radius. The measurements are
generated from manoeuvring tests which are conducted on pavement.
However, the pavement is not completely flat, causing the cargo module to be
slightly unlevel at times and influencing the measurement. The hysteresis
present is also not taken into account.
75/107
Thanks to ability to tilt, the driver only experiences a fraction of the lateral
acceleration compared to a fixed frame vehicle. While testing the maximum
possible velocity driving a circle with a radius of 4 m the driver only
experienced a lateral acceleration of 0.89 m/s2 while a fixed vehicle
experienced 4.5 m/s2 [9]. The driver experiences less lateral acceleration with
the concept vehicle. A comfortable lateral acceleration is stated to be below
0.9 m/s2 [122].
Due to the low torque the torsion system has at near-vertical positions, the
improvement in low-speed stability is minimal. Just like with a conventional
bicycle, extra concentration is required to cycle at low velocity compared to
high velocity. The cyclist experiences a minimal difference with or without
the torsion system at low velocity. Cycling above 10 km/h, the vehicle has
self-stable properties and does not require the aid of a torsion system to remain
upright. Increasing the torque stiffness can improve the low-speed stability
and limit the parking tilting, but can limit the ability to tilt while turning. On
a traditional bicycle, while cycling a straight line at 9 km/h, the roll angle
remains between −1.5° and 1.5° [123]. To remain stable, the cyclist has to
balance by tilting his torso or steering. To aid in the low-speed stability, the
torsion system counter-torque should be equal or higher to compensate for the
imbalance.
The torque created by the imbalance is calculated using equation 4. Centripetal
force is neglected as the stability in a straight line is considered. A rigid cyclist
body is considered, meaning the cyclist does not lean in the opposite direction
to aid in the balance. At a tilt angle of 1.5°, the driver module with the cyclist
generates a torque of 24.39 Nm.
When the torsion system is 10 times stiffer, the counter-torque is higher as can
be seen in figure 52 and should improve the low-speed stability. This cannot
be tested with the current prototype due to the difficulty of changing the
torsion system. This stiffer torsion system reduces the tilting when parked but
also reduces the ability to tilt while cornering. These changes can be calculated
as before.
76/107
Figure 52 Graph of the current and 10 times stiffer torsion system. These
are compared with the torsion the driver module with and without cyclist
creates on level surface and inclined (8%).
The changes while tilting due to the stiffer torsion system are calculated using
equation 6. These changes can be seen in figure 53.
Figure 53 Tilting calculation of torsion system and 10 times stiffer torsion
system and difference due to torsion system.
77/107
The stiffer torsion decreases the tilting according to the velocity. The influence
is not shown over the complete turning radius as the vehicle’s velocity
increases the minimum possible turning radius. During a big turning radius,
the stiffer torsion spring causes little tilting reduction. This is a minor change
and has a limited influence on the manoeuvrability. However, the influence of
the torsion spring increases when the velocity is increased or at a smaller
turning radius. In extreme cases, for example when cycling at 25 km/h through
a 10 m turning radius, this can reduce the tilting angle by 8°. This results in an
increased lateral acceleration of 1.5 m/s2 which is still considered as a normal
driving acceleration [122]. While cycling normally, the stiffer torsion spring
would only account to a tilting reduction of 2°.
5. Discussion
The torsion system with an integrated angle limit is an optimal choice. The
torsion system meets the requirements. It prevents the driver module from
tipping over. The system can easily and cheaply be integrated into a hinge and
does not require any control/adjustment when starting/stopping unlike a
locking mechanism or a side stand. Furthermore, the torsion system improves
the handling qualities. The damping properties of the system reduce vibrations
in the vehicle. Unfortunately, the driver module can still tilt 22° while parked
causing the steer to extend 30 cm outside the vehicle’s width. The cyclist has
to straighten the driver module every time he gets on the vehicle. Furthermore,
the low torsion of the current torsion system does not aid in the low-speed
stability. An increase in the stiffness of the torsion system results in a tilting
angle reduction. The titling reduction during normal driving conditions is only
2 degrees.
This indicates the stiffer torsion system will improve the vehicle’s ergonomics
further, but practical tests are required to find an optimal balance between low-
speed stability and high-velocity manoeuvring.
6. Conclusion
A torsion system is implemented in the concept vehicle. The system meets all
the requirements. It prevents the driver module from tipping over when
parked. The driver module can still tilt 17° but will not fall over. The torsion
system has a small effect while cornering. The vehicle can still tilt while
driving. It does not interfere with the handling qualities of the vehicle The
system even provides damping reducing vibrations while cycling. Although
the torsion system provides enough counter-torque the module doesn’t fall
over, an angle limit is still incorporated to ensure a maximum tilting angle in
extreme cases. A stand is added to the cargo module to aid while loading and
unloading.
78/107
Chapter 5: Valorisation plan
During the course of the PhD, a novel Last-Mile Vehicle (LMV) has been
developed. The prototype is a ‘proof of concept’ that complies with previously
stated desired requirements [9]. The prototype has proper handling qualities
independent of the cargo [69]. The developed drivetrain is as intuitive to use
just as a conventional e-bike and ensures pedal comfort [80]. The LMV has a
higher payload capacity (euro pallet sized cargo up to 200 kg) compared to the
majority of current cargo bikes. Furthermore custom components are
developed which can be sold separately to use incorporate in novel specialized
cargo bikes.
1. Market research
LMV is primarily developed for last-mile delivery in city centres to transport
multiple small parcels over a short distance. The e-commerce delivery markets
and the opportunities to launch LMV are reviewed. Competition of current
traditional vans and the similar vehicles are compared. Since the LMV is a
prototype, adjustments are needed to create a market-ready model. The
possible and required improvements are described and a cost estimate is made.
Furthermore, the research led to the development of custom parts in
cooperation with manufacturing companies. These components are integrated
into the LMV and can be utilized in other novel industrial cargo bikes.
1.1. Courier, Express and Parcel services
LMV is developed especially for Courier, Express and Parcel (CEP) services.
CEP services can are divided in three segments. B2B, B2C and C2C are the
different segments of which B2C accounts to 56% [124]. Although CEP
service is currently dominated by van delivery, the use of cargo bikes has
increased substantially. Around 1% of e-commerce parcels are currently
delivered using cargo bikes [125]. The transition towards cargo bikes has
already begun. 87% of the CEP companies already use cargo bikes for
deliveries. The average number of cargo bikes has more than quadrupled from
2019 to 2022. Furthermore CEP market forecasts a 10% Compound Annual
Growth Rate (CAGR). In 2021, the adoption of cargo bikes in the CEP service
accounted for 43 % of the total cargo bike market sales [126].
79/107
The growing use of cargo bikes is due to several factors.
• The total volume of CEP services has quadrupled in Belgium last
decade. This is an increase from 72 million delivered units in 2010
to 336 million units in 2020 [127]. The increased volume requires
the CEP services to optimize and/or increase their fleet to be able to
handle the increased parcel volume.
• In five years, e-commerce has grown from 10% to 20.4% of the
global retail sales [128]. E-commerce is the main driver of the CEP
market growth. The increasingly competitive e-commerce market
drives the CEP market to reduce delivery cost. Cargo bikes are ideal
to optimise the parcel fleet. 85% of all parcels can be delivered by
cargo bikes [117, 129]. Cargo bikes are more cost-effective than
vans [130]. They have a much lower total cost of ownership and can
deliver 60% faster in city centres.
• Recent substantial growth of e-commerce can be attributed to Covid-
19. During to the lockdown, there was a year-over-year increase of
17.5% of online stores in 2021 [131]. This sudden growth remains
after the restrictions with and forecasts speculating a CAGR of 7.5%
during 2023-2028 [132].
• Governments are encouraging companies to decrease pollution. This
is done by investing in Low-Emission zones, shunning traffic from
city-centres, improving bicycle infrastructure and incentivising
companies to use less polluting alternatives. Electric cargo bikes can
cut carbon emissions by 90% compared to diesel vans accounting
production, materials and battery use [133].
• CEP companies aim to obtain a zero-emission fleet. This can be done
by using electric vans or cargo bikes. Due to the lighter mass and
mechanical propulsion, electric cargo bikes pollute less than 10%
compared to electric vans while driving [134].
• Electric cargo bikes have a 60% lower total cost of ownership
compared to a van and can have 10 times lower expenses per parcel
thanks to lower acquisition price, lower maintenance, fuel/energy
[31, 32, 135].
• Delivery using cargo bikes can deliver more efficient compared to
vans due to less congestion problems, more parking possibilities and
fewer restricted driving areas [136]. Depending on the region this
can be about 60% faster.
1.2. Pickup points
An alternative way to optimise last-mile delivery is using pickup points.
Pickup points can be in-store services or self-service parcel lockers. Pickup
points combine multiple drop locations. This results in a more efficient
80/107
last-mile. They are becoming an increasingly popular option for e-commerce
deliveries, particularly in urban areas. Parcel lockers are secure, automated
storage units that allow customers to pick up their packages at a convenient
time and location. According to a report by Pitney Bowes, parcel lockers are
used for 9% of e-commerce deliveries in Europe and this number is expected
to grow in the coming years.
Even though 69% live within a 1 km radius of a collection point in Belgium,
40% of pickup points are company-specific pickup points. Furthermore, parcel
locker only work if customers are willing to retrieve their parcel at a locker.
The majority of the customers still prefer the convenience of home delivery.
70% is still delivered at home [137, 138]. If home delivery costs 3 euro more
than a pickup point, 50% still prefer home delivery [5].
The use of parcel lockers can reduce the number of failed deliveries, as well
as the need for vans to make multiple delivery attempts. While the popularity
of pickup points increases, this doesn’t particularly mean a bad thing since the
parcels can be delivered using cargo bikes.
1.3. Other uses
87% of businesses with cargo bikes is used for delivering goods [139]. Apart
from CEP services, cargo bikes have other purposes. 5% is used to supply
services instead of parcels. For example, businesses that require limited tools
like breakdown service, gardening service, cleaning service, electrician,
plumber, etc. Furthermore, 5% of businesses use cargo bikes as a taxi. The
design of the cargo bike allows for custom add-ons depending on the cargo
bike’s use.
1.4. Bicycle market
The global bicycle market was valued at USD 59.33 billion in 2021 and is
expected to have a Compound Annual Growth Rate of 8.2% from 2022 to
2030 [140]. In 2019 electric cargo bikes accounted for 1.3% of the bicycle
market share [141]. Between 2016 and 2020 the electric cargo bike market
had a 7.7% annual grow rate. The expected electric cargo bike market growth
ranges from 12.7% to 34.8% CAGR and is expected to reach a market revenue
ranging between USD 1.9 and 3.8 billion [142]. As stated before, The present
surge in growth can be attributed, at least in part, to the COVID-19 pandemic
and the consequent growth of e-commerce.
1.5. Region
The transition towards cargo bikes and sustainable alternatives for last-mile
delivery is prominently happening in Europe. Due to the development of
ancient cities, Europe is better suited for cargo bike delivery. Furthermore,
Europe has a stronger push towards sustainability and environmentally
81/107
friendly incentivising CEP companies to shift towards cargo bikes. Due to a
higher percentage of cyclist in Europe, there is also a better cycling
infrastructure present. Cycle logistics calculated 51% off all motorised trips in
an average European city are related to light goods transport (cargo below 200
kg) and could be shifted to bicycle transport [14].
Europe is the main market for electric cargo bikes accounting for 68.3% of the
global market share. An estimated 400.000-500.000 electric cargo bikes will
be sold in 2022 in Europe [46, 143]. Cargo bikes can be categorised as a two-
wheeled, three-wheeled or four-wheeled bike. With 49% of the electric cargo
bike market share, three-wheeled cargo bikes dominated the market in 2021
with a revenue of USD 0.58 billion [144].
In 2021, 30% of Belgian revenue in companies is generated by e-commerce.
Only Ireland and Chechia have a higher percentage [145]. During 2022, e-
cargo bikes sales in Belgium had a year over year increase of 760% [146].
2. Important design specifications
A lot of different cargo models already exist for a light-medium cargo
capacity. They have a cargo volume of around 0.5 m³ and are mostly
developed for commercial use. They have a maximum cargo capacity of
150 kg.
At the start of the development European Regulations limited the maximum
width of a bicycle (L1e-A) to 1 m [147]. Individual countries can exclude
those limitations to their own. For example, in Belgium and the Netherlands
single track vehicles have a maximum width of 0.75 m and multitrack vehicles
have a maximum width of 2.5 m and 1.5 m respectively [37, 148]. The
maximum width of 2.5 m is extremely wide and is stated in regards of multi-
person go-carts or a so called beer bike. Trailers have a maximum allowed
width of 1 m in general and 1.2 m for permitted pilot projects [38]. However
vehicles wider than 1 m cannot utilise bicycle lanes in Belgium. Furthermore
the bicycle infrastructure is not everywhere adapted to cargo bikes. Thus,
narrower cargo bike are easier to use bicycle infrastructure.
3. Competitors
At the start of the doctoral research, no cargo bikes were commercially
available with a similar payload capacity and handling qualities. The most
commonly used type were front loader bicycles like the Urban Arrow or the
Bullitt [49]. While developing, multiple tilting delta tricycles have appeared
with an increased cargo capacity. This proves the gap in the market and shows
the need for this type of cargo bike.
A summary of available cargo bikes is made:
82/107
Urban Arrow
Price
€ 7.300
Cargo
capacity:
0.33 m³
90 cm x
60 cm
125 kg
Size: l x w
(cm)
274 x 70
Drivetrain:
Mechanical
driven with
250 W mid-
drive
Classification
EPAC
https://urbanarrow.com/nl/
The electrified Bullitt
Price
€ 5.000
Cargo
capacity:
0.20-0.40 m³
70 cm x
40 cm
125 kg
Size: l x w
(cm)
243 x 46
Drivetrain:
Mechanical
driven with
250 W mid-
drive
classification
EPAC
https://www.larryvsharry.com/
Fulpra
Price
€ 14.990
Cargo
capacity:
3.00 m³
160 cm x
80 cm
125 kg
Size: l x w
(cm)
325 x 100
Drivetrain:
Mechanical
driven with
1000 W
mid-drive
Classification
L1e-A
https://fulpra.com/en/
83/107
Cargo chariot
Price
€ 13.990
Cargo
capacity:
1.50 m³
120 cm x
60 cm
275 kg
Size: l x w
(cm)
240 x 100
Drivetrain:
Mechanical
driven with
250 W
mid-drive
classification
EPAC
https://www.cargocycling.com/
gleam
Price
€ 8.250
Cargo
capacity:
1.50 m³
120 cm x
60 cm
275 kg
Size: l x w
(cm)
240 x 100
Drivetrain:
Mechanical
driven with
250 W mid-
drive
classification
EPAC
https://gleam-bikes.com/
VOK XL
Price
€ 13.499
Cargo
capacity:
2.00 m³
160 cm x
85 cm x
135 cm
200 kg
Size: l x w
(cm)
310 x 90
Drivetrain:
Series hybrid
drive system
with
hub-motors
classification
EPAC
https://vokbikes.com/vok-bike-xl/
84/107
YakBike
Price
€ 17.500
Cargo
capacity:
2.40 m³
170 cm x
85 cm x
125 cm
250 kg
Size: l x w
(cm)
280 x 110
Drivetrain:
Series hybrid
drive system
with
hub- motors
classification
EPAC
https://www.yakbike.co/en/
LMV
Price
not on market
Cargo
capacity:
1.50 m³
125 cm x
82 cm
200 kg
Size: l x w
(cm)
240X100
Drivetrain:
Series hybrid
drive system
with
hub-motors
classification
25 km/h
https://iiw.kuleuven.be/onderzoek/itl
As seen above, There is already a variety of cargo 100+ kg payload capacity.
The competitors have similar specifications as the LMV. Only the Fulpra and
the VOX XL also have a minimum cargo area width of 80 cm combined with
a maximum e-cargo bike width of 1 meter. However, VOX XL uses a cargo
area above the wheels instead of between them and is not equipped to transport
euro pallets. With a worldwide yearly sale of 500.000 e-cargo bikes and rising
electric cargo bikes in Europe there is still room for competitors [46].
85/107
4. Product features (unique selling points)
Although there are other cargo bikes available with sizable cargo capacities,
the LMV stands out due to its array of distinctive features. These unique
attributes set it apart from competitors, making it a standout choice in the
market.
4.1. Cargo capacity
LMV has the ability to transport euro pallets (120 cm x 80 cm). This can be
un/loaded using a pallet jack and a ramp.
4.2. Handling qualities
Thanks to the tilting system, the LMV has a comfortable driving feeling and
proper handling qualities. The geometry of the vehicle allows for a short
turning radius which is ideal for narrow city centre streets. The cargo area is
located between the rear-wheel creating a low centre of gravity. The low
centre of gravity creates a stable vehicle.
4.3. Configuration
The load platform consist of a standard flat-bed, but can be configured
according to the customers’ requirements. The low load platform allows for
easy loading/unloading of a euro pallet or a city container. The design allows
tire replacement without needing to jack or unload the vehicle facilitating
maintenance and repair.
4.4. Drivetrain
The vehicle uses a series hybrid drivetrain. In a series hybrid drivetrain,
propulsion is completely done by hub motors and has no mechanical
transmission with the pedals. The lack of mechanical components ensure a
maintenance friendly vehicle. Furthermore, the vehicle uses regenerative
braking, which reduces the wear of brakes and increases the vehicle’s range.
The integrated CVT lets the cyclist pedal at a comfortable cadence and allows
the vehicle to pedal charge while stationary. While parting on a conventional
e-bike can require a high pedal torque especially with hub-motors that
generally use a rotary pedal sensor.
An electric differential or torque vectoring is implemented improving the
stability and control while manoeuvring.
This drivetrain can also be sold as a module and can be modularised with one
or more hub motors. It can also be implemented with multiple pedal generators
on tandems. Interest in the drivetrain has already be shown by parcel couriers
and bicycle manufacturers.
86/107
Unique selling points are summarized below:
• tilting system ensures comfortable driving feeling and proper
handling qualities
• a configurable cargo hold
• euro pallet sized cargo hold
• regenerative braking
• continuous variable transmission
• pedal charging when parked
• low maintenance
• maintenance friendly (flat tire can be replaced on a fully loaded
vehicle)
• bicycle lane allowed
5. Cost estimates
The sheet metal and rivet assembly allowed for a fast mechanical assembly of
only 4 hours during prototype assembly. Electrical assembly took one week
due to finetuning and adjustments. Mechanical assembly can still be reduced
when using specialised assembly tools. Electronic assembly was a labour
intensive process due to soldering of multiple custom PCB’s and adaptations
mid-assembly. Optimising the electrical design will simplify the assembly.
The prototype had a € 4.000 material cost. The average material cost is
estimated to drop 25 % for series production. Optimising the design further
can further reduce the material cost. For example, the cargo module is a sheet-
metal construction which can be designed using thinner sheet metal reducing
the weight. This would result in a € 3.000 material cost. A retail price of €
8.000 results in a feasible selling price and also corresponds with the
competitors cargo bike market price.
6. Required improvements
Although the prototype proves the capabilities of the concept, it is not yet
ready as a product. The prototype still requires some adaptations for the
vehicle to be ready for series-production. Industrial couriers often cycle
intensively and overload the vehicle. Currently the maximum cargo capacity
is limited by the wheels. Spoke-less flanged wheels are an alternative that can
increase the cargo capacity.
Improvements in the design can also be made to reduce manufacturing cost.
Currently, the frame primarily consist of a riveted sheet metal structure
already eliminating the need for costly welding. This structure can be
optimized to limit the used material and reduce the vehicle weight.
87/107
The tilting system was created using an expensive slewing bearing. In
corporation with Beluma, a module was designed to replace the slewing
bearing and already incorporates a torsion spring for improved handling
qualities. Furthermore, the design can be finetuned to further reduce the
manufacturing costs.
For the prototype, the electronics of the vehicle currently consist of multiple
PCB’s soldered together. Manufacturing of the current electronics are labour
intensive and expensive. A novel single PCB requires no manual soldering
and added connections allow for easy electrical wiring.
7. Selling components
A different direction that is possible is to sell components that were developed
during the doctoral research of the LMV. One of these components is the pedal
series drivetrain. A control loop is developed to simulate pedalling with a
conventional e-bike and can work on any BLDC-motor [80].
This drivetrain is highly modular and can be integrated in conventional
bicycles or custom vehicles. The electrical transmission reduces design
constraints and expands design possibilities. The drivetrain can work in
multiple configurations. Propulsion can be done using one or multiple hub
motors. Furthermore the pedal generator can be used on a single driver vehicle
like the LMV or a multi-driver vehicle like a tandem which allows each driver
to pedal at their own comfort. Furthermore, the generator can also be hand
driven for people with disabilities. An integrated Continuous Variable
Transmission can automatically set a comfortable gear ratio and resistance on
the pedal generator. This can lower the required pedalling force to depart
which is beneficial for elderly people.
The EU Commission has recently excluded Series Hybrid Cycles from L-
category and classified as an EPAC. This facilitates the homologation and aids
further technological development for Series Hybrid Cycles [149].
In 2021 Schaeffler has developed the “free drive” in cooperation with
Heinzman [95]. Those big companies further lobby to improve/clarify the
regulations to facilitate the use of Series Hybrid. The only other known
commercially available Series Hybrid module is “PMD” from Mando [94].
No price for the “free drive” module is yet available however during selection
of the hub motors a Heinzman hub motor had a 4 times higher cost compared
to the hub motors used in the LMV while having similar specifications.
88/107
Chapter 6: Conclusion and future work
The research aims to develop a high-performance cargo bicycle as a more
sustainable alternative to the traditional courier vans used for last-mile
logistics. This is one of the many solutions to optimize the last-mile delivery.
Cargo bikes are an eco-friendly alternative as they cut the total carbon
emissions of a cargo bike by 90% compared to diesel vans. Furthermore, cargo
bikes can optimize the last-mile delivery as they have less delays due to
congestions. They can optimize their route by using low-emission zones and
bicycle lanes and have less parking problems.
Multiple interviews with different stakeholders reveals the shortcomings of
the current cargo bikes and the need for a novel cargo bike. Stakeholders
emphasize the importance of enhancing the cargo capacity and the handling
qualities in a novel last-mile vehicle.
These insights manifest in the design of a tricycle featuring a level cargo
module connected to a tilting driver module. This design allows the driver to
lean during manoeuvres, similar as with conventional bicycles, while the
cargo module remains stable and level. The hinge mechanism between the
driver and cargo module plays a pivotal role in decoupling the cargo weight
from the driver’s movement. This concept combines the desired high cargo
capacity while maintaining a manoeuvrable vehicle.
A prototype is developed to determine the effect of different hinge
orientations. Manoeuvres are executed using multiple configurations of the
hinge’s position while transporting different cargo weight. Those tests note a
less stable cargo bike with a hinge mounted on a higher position. The higher
hinge position causes the drivers module to tilt at a higher position while
leaning, giving the driver the feeling the cargo bike is slipping away
underneath them. Modifying the angle of the hinge has a bigger influence.
Tilting the hinge downward to the driver module has the benefit that the cargo
bike can make shorter turns making it more manoeuvrable. A more
manoeuvrable cargo bike has the downside of being less stable. Modifying the
hinge to an upward tilt towards the driver had the opposite effect. The cargo
bike’s turning radius was increased, making the manoeuvres more difficult.
However the bike feels more stable.
The configurations always have the same effect, independent of the cargo
bike’s payload. A slightly less stable, but extra manoeuvrable configuration is
opted for the second prototype.
89/107
The cargo bike’s drivetrain, a pedal series hybrid drivetrain is implemented in
the second prototype. This innovative drivetrain integrates a pedal generator
to harness human power and an electrical propulsion system for vehicle
propulsion. This electrical drivetrain reduces the mechanical complexity and
has less design restrictions for the cargo bike. Furthermore, this type of
drivetrain offers notable advantages such as independent torque control,
regenerative braking with integrated ABS, variable gearing and the ability to
park in reverse.
Ensuring a natural pedal feeling requires the development of a control system.
To mimic the dynamics of traditional biking, the pedal resistance must vary
according to the pedal crank orientation. This control system regulates the
resistance of the pedal generator dependent of the pedal rotational velocity,
the wheels rotational velocity and the exerted wheel power, achieving a
maximum resistance during a horizontal pedal crank orientation and a
minimum resistance at a vertical pedal crank orientation.
To validate the effectiveness of the pedal generator, data gathered from a
measuring bicycle is compared to the pedal generators data. This concludes in
consistent resistance variations according to the crank orientation. The cargo
bike exhibits a higher variation in the instantaneous angular velocity due to
the electrical gearing which allows for slight slipping in accordance with the
wheel’s gearing.
To further improve the cargo bike’s handling qualities and ergonomics, a
torsion in the hinge between the driver and cargo module is implemented. This
enhancement prevents the driver module from tipping while parked and also
improves the low-speed stability. The influence of this torsion is further
examined for tilting during turns and a minimal effect was calculated.
While the developed prototype achieves a commendable balance between
cargo capacity and handling qualities, it is crucial to acknowledge that further
development is needed to transform this prototype into a production-ready
vehicle.
Future work may entail conducting user experience studies to offer deeper
insights in the practicality and usability of the novel cargo bike. Gathering
feedback can guide future refinements. As the bike industry keeps evolving,
similar commercial cargo bikes to the novel cargo bike have emerged.
Examples of this are the Fulpra and yakbike. Comparing the handling qualities
can provide a comprehensive understanding of the novel cargo bike’s handling
qualities.
90/107
Bibliography
[1] C. Tejada and A. Conway, "What Happens Before the Last
Mile? Exploring a Package’s Journey," Transportation
Research Record, vol. 2677, no. 2, pp. 15-32, 2023/02/01
2022, doi: 10.1177/03611981221128804.
[2] M. I. Statista. "B2C E-commerce." Statista Market Insights.
https://www.statista.com/outlook/dmo/ecommerce/belgiu
m (accessed March, 2023).
[3] S. Dolan. "The challenges of last mile delivery logistics and the
tech solutions cutting costs in the final mile."
https://www.insiderintelligence.com/insights/last-mile-
delivery-shipping-explained/ (accessed April, 2023).
[4] DispatchTrack. "Breaking Down Last Mile Delivery Costs."
DispatchTrack. https://www.dispatchtrack.com/blog/last-
mile-delivery-costs-breakdown (accessed April, 2023).
[5] M. Joerss, J. Schröder, F. Neuhaus, C. Klink, and F. Mann,
"Parcel delivery The future of last mile," McKinsey&Company,
2016. [Online]. Available:
https://www.mckinsey.com/~/media/mckinsey/industries/tr
avel%20logistics%20and%20infrastructure/our%20insights/h
ow%20customer%20demands%20are%20reshaping%20last
%20mile%20delivery/parcel_delivery_the_future_of_last_mi
le.pdf
[6] FarEye. "What are last-mile delivery costs? 7 ways to reduce
them and boost last-mile delivery profits." FarEye.
https://fareye.com/resources/blogs/last-mile-delivery-costs
(accessed July, 2023).
[7] P. Dizikes. "The 6-percent solution." MIT News Office.
https://news.mit.edu/2010/corporate-greenhouse-gas-1108
(accessed June, 2023).
[8] S. Suzan, "Easy Ride: why the EU ttruck CO2 targets are unfit
for the 2020s," Transport & Environment, 2021.
[9] J. D'hondt, M. Juwet, E. Demeester, and P. Slaets,
"Development of an electric tricycle for service companies
and last-mile parcel delivery," vol. 17, no. 2, 2022, doi:
10.20858/tp.2022.17.2.15.
91/107
[10] K. Reiter, S. Wrighton, and R. Rzewnicki, "Potential to shift
goods transport from cars to bicycles in European cities," in
Moving Europe forward, ed, 2014.
[11] (2020). European Sustainable and Smart Mobility Strategy.
[12] L. Caggiani, A. Colovic, L. P. Prencipe, and M. Ottomanelli, "A
green logistics solution for last-mile deliveries considering e-
vans and e-cargo bikes," Transportation Research Procedia,
vol. 52, pp. 75-82, 2021/01/01/ 2021, doi:
https://doi.org/10.1016/j.trpro.2021.01.010.
[13] E. Verlinghieri, I. Itova, N. Collignon, and R. Aldred, "The
Promise of Low-Carbon Freight," Westminster, 2021.
[14] S. Wrighton and K. Reiter, "CycleLogistics – Moving Europe
Forward!," Transportation Research Procedia, vol. 12, pp.
950-958, 2016/01/01/ 2016, doi:
https://doi.org/10.1016/j.trpro.2016.02.046.
[15] C. Giglio, R. Musmanno, and R. Palmieri, "Cycle logistics
projects in Europe: Intertwining bike-related success factors
and region-specific public policies with economic results,"
Applied sciences, vol. 11, no. 4, pp. 1-31, 2021, doi:
10.3390/app11041578.
[16] F. Arnold, I. Cardenas, K. Sörensen, and W. Dewulf,
"Simulation of B2C e-commerce distribution in Antwerp using
cargo bikes and delivery points," European Transport
Research Review, vol. 10, no. 1, p. 2, 2017/12/04 2017, doi:
10.1007/s12544-017-0272-6.
[17] D. Nowak, "How sustainable is the Pedal Me fleet?," ed: Pedal
Me, 2021.
[18] M. Mucowska, "Trends of Environmentally Sustainable
Solutions of Urban Last-Mile Deliveries on the E-Commerce
Market—A Literature Review," Sustainability, vol. 13, no. 11,
2021, doi: 10.3390/su13115894.
[19] J. Clement. Coronavirus impact on retail e-commerce website
traffic worldwide as of June 2020, by average monthly visits
[Online] Available:
https://www.statista.com/statistics/111259
[20] R. Villa and A. Monzón, "Mobility Restrictions and E-
Commerce: Holistic Balance in Madrid Centre during COVID-
19 Lockdown," Economies, vol. 9, no. 2, 2021, doi:
10.3390/economies9020057.
92/107
[21] I. Cárdenas, J. Beckers, and T. Vanelslander, "E-commerce
last-mile in Belgium: Developing an external cost delivery
index," Research in transportation business & management,
vol. 24, pp. 123-129, 2017, doi: 10.1016/j.rtbm.2017.07.006.
[22] J. Maes, "The potential of cargo bicycle transport as a
sustainable solution for urban logistics," Faculteit
Bedrijfswetenschappen en Economie, Univversiteit
Antwerpen, 2017.
[23] J. Gruber, A. Kihm, and B. Lenz, "A new vehicle for urban
freight? An ex-ante evaluation of electric cargo bikes in
courier services," Research in transportation business &
management, vol. 11, pp. 53-62, 2014, doi:
10.1016/j.rtbm.2014.03.004.
[24] W. Kingdon, M. Figliozzi, L. Kingdon, and A. Wilkitzki, "Analysis
of freight tours in a congested urban area using disaggregated
data: characteristics and data collection challenges," 12/15
2007.
[25] "On average, how many packages and stops do you have a
day?" Reddit.
https://www.reddit.com/r/AmazonDSPDrivers/comments/lt
9y9t/on_average_how_many_packages_and_stops_do_you/
(accessed June, 2023).
[26] J. Allen et al., "Understanding the impact of e-commerce on
last-mile light goods vehicle activity in urban areas: The case
of London," Transportation Research Part D: Transport and
Environment, vol. 61, pp. 325-338, 2018/06/01/ 2018, doi:
https://doi.org/10.1016/j.trd.2017.07.020.
[27] B. Straight. "‘Stretched to its limits’: Survey finds last-mile
driver workforce is pushing boundaries." Freight waves.
https://www.freightwaves.com/news/last-mile-delivery-
drivers-being-asked-to-deliver-more-packages (accessed
June, 2023).
[28] L. Prejean. "Vehicle Trends in Last-Mile Delivery." Worktruck.
https://www.worktruckonline.com/10164139/vehicle-
trends-in-last-mile-delivery (accessed June, 2023).
[29] H. Winklbauer, "cross-border e-commerce shopper survey
2021 ", ed, 2021.
[30] M. Garland. "Pitney Bowes wants heavier parcels in its
network to boost profits." Supplychaindive.
93/107
https://www.supplychaindive.com/news/pitney-bowes-
pursues-heavier-parcels-q4-earnings/642560/ (accessed
August, 2023).
[31] J. ERLANDSSON. "The Armadillo electric cargo bike use 6 % of
the electricity of a small electric van."
https://www.velove.se/news/armadillo-cargo-bike-use-6-
electricity-small-electric-van (accessed March, 2023).
[32] K. Ai, "Data-driven Evaluation of Cargo Bike Delivery
Performance in Brussels," 2023.
[33] J. Palau, "Environmental impact and cost price analysis of
electric cargo bikes and electric vans for freight transport. ,"
Master in management engineering, with a focus on Supply
chain Management and Business Analytics, University of
Liège, 2021.
[34] T. S. Buch and P. Greibe, "Bredde af cykelstier: Analyse
af adfærd og kapacitet," 2014. [Online]. Available:
https://www.trafitec.dk/sites/default/files/publications/Bre
dde%20af%20cykelstier_Analyse%20af%20adf%C3%A6rd%2
0og%20kapacitet.pdf
[35] T. E. PARLIAMENT and COUNCIL, the approval and market
surveillance of two- or three-wheel vehicles and quadricycles,
2020.
[36] A. Roetynck, "Electric Cargo Bikes: Technical Rules &
Regulations," in ECLF conference, Vienna, 2017, 2017 ed.
[Online]. Available:
https://eclf.bike/presentations17/D1%20Annick%20Roetync
k%20ECLF%202017.pdf
[37] R. de Groot, Ontwerpwijzer fietsverkeer. Ede : CROW, 2016.
[38] (1975). Koninklijk besluit houdende algemeen reglement op
de politie van het wegverkeer en van het gebruik van de
openbare weg.
[39] m. vlaanderen, "ONTWERPICHTLJINEN VOOR
FIETSVOORZIENINGEN," ed, 2017.
[40] (2018). Vademecum fietsvoorzieningen - Minimumbreedte
fietspaden.
[41] (2000). Danish Guidelines for planning and design, Danish
Road Directorate.
[42] P. Greibe and T. S. Buch, "Capacity and Behaviour on One-way
Cycle Tracks of Different Widths," Transportation Research
94/107
Procedia, vol. 15, pp. 122-136, 2016/01/01/ 2016, doi:
https://doi.org/10.1016/j.trpro.2016.06.011.
[43] G. Fransen, "Van bakfietsen tot snelle elektrische fietsen: wie
mag er wel op het fietspad en wie niet?," ed: Het Nieuwsblad,
2020.
[44] A. Behrensen and A. Sumer, "European Cargo Bike Industry
Survey Results," City Changer Cargo Bike, 2020. Accessed:
25/04/2022. [Online]. Available:
https://www.cargobike.jetzt/europaeische-marktumfrage-
2020/#Ergebnisse
[45] S. Wens and F. Vlaanderen, "Fiets DNA," 2018. [Online].
Available: https://fietsberaad.be/wp-
content/uploads/FietsDNA_2018_A5_lwr.pdf
[46] A. Behrensen, " European Cargo Bike Industry Survey,"
Cyclelogistics, 2021.
[47] G. Farman, "Cargo Bike Analysis and Field Testing," Topsector
Logistiek, 2020. [Online]. Available:
https://www.topsectorlogistiek.nl/wp-
content/uploads/2020/11/Cargo-bike-HR-testing-
method.pdf
[48] A. Behrensen and A. Sumer, "First European Cargo Bike
Industry Survey Results," City Changer Cargo Bike, 2020.
Accessed: 25/04/2022. [Online]. Available:
https://www.cargobike.jetzt/europaeische-marktumfrage-
2020/#Ergebnisse
[49] S. Narayanan, J. Gruber, G. Liedtke, and C. Antoniou,
"Purchase intention and actual purchase of cargo cycles:
Influencing factors and policy insights," Transportation
Research Part A: Policy and Practice, vol. 155, pp. 31-45, 01/01
2022, doi: 10.1016/j.tra.2021.10.007.
[50] L. Blazejewski, G. Sherriff, and N. Davies, Delivering The Last
Mile Scoping the Potential for E-cargo Bikes Healthy Active
Cities. 2020.
[51] B. Lenz and E. Riehle, "Bikes for urban freight?,"
Transportation research record, no. 2379, pp. 39-45, 2013,
doi: 10.3141/2379-05.
[52] S. Clements, "Cycle Logistics Study," 2019. [Online]. Available:
https://crossriverpartnership.org/wp-
95/107
content/uploads/2019/03/20190520_Element-
Energy_Cycling-logistics-study_FINAL-REPORT-1.pdf
[53] W. P. van Amstel "DHL Netherlands presents their new cargo
bike: the Chariot," ed. City Logistics, 2021.
[54] UPS, "Lessons from our e-bike journey," ed, 2020.
[55] FED-ex, "FedEx Express Continues Journey Towards Zero
Emissions Delivery, as Edinburgh, Glasgow and Cambridge
Become the Next UK Cities to Welcome E-cargo Bikes," ed,
2021.
[56] J. Gruber and S. Narayanan, "Travel Time Differences between
Cargo Cycles and Cars in Commercial Transport Operations,"
Transportation research record, vol. 2673, no. 8, pp. 623-637,
2019, doi: 10.1177/0361198119843088.
[57] Ziegler, "A revolutionary, eco-friendly freight solution for
Brussels.," ed: Ziegler, 2022.
[58] U. Arrow. "The Tender will handle everything you throw at it."
https://urbanarrow.com/business-bikes/tender/tender-
2500/ (accessed april, 2022).
[59] E. Brust, "Zweirad-Industrie-Verband (ZIV)
Wirtschaftspressekonferenz," Berlin, 2021, Berlin.
[60] D. Maritan, Practical Manual of Quality Function Deployment.
Cham: Springer International Publishing AG, 2014.
[61] (2017). Cycles- Electrically powere assisted cycles - EPAC
Bicycles EN 15194:2017.
[62] UNECE, "Uniform provisions concerning the approval of
vehicles of categories L1, L2, L3, L4 and L5 with regard to
braking," in Regulation No 78 of the Economic Commission for
Europe of the United Nations (UNECE), ed. official Journal of
the European Union, 2015.
[63] M. Bahrami, H. Mokhtari, and A. Dindar, "Energy regeneration
technique for electric vehicles driven by a brushless DC
motor," IET Power Electronics, vol. 12, no. 13, pp. 3397-3402,
2019/11/01 2019, doi: https://doi.org/10.1049/iet-
pel.2019.0024.
[64] E. Tournon, P. Venet, B. Barbedette, A. Lelievre, J. Aubry, and
A. Sari, "Efficiency Comparison between Series Hybrid Bike
and Traditional Bike," in Fourteenth International Conference
on Ecological Vehicles and Renewable Energies (EVER),
96/107
Monaco · Monte-Carlo, 2019: IEEE, pp. 1-7, doi:
10.1109/EVER.2019.8813618.
[65] K. Zawirski, K. Nowopolski, B. Wicher, D. Janiszewski, B.
Fabiański, and K. Siembab, "Gearless Pedaling Electric Driven
Tricycle," vol. 324, ed: Cham: Springer International
Publishing, 2014, pp. 411-421.
[66] C. Song, Y. Shao, S. Song, S. Peng, and F. Xiao, "A Novel Electric
Bicycle Battery Monitoring System Based on Android Client,"
Journal of engineering (Cairo, Egypt), vol. 2017, pp. 1-11,
2017, doi: 10.1155/2017/2579084.
[67] O. Ustun, G. Tanc, O. C. Kivanc, and G. Tosun, "In pursuit of
proper BLDC motor design for electric bicycles," in 2016 XXII
International Conference on Electrical Machines (ICEM), 4-7
Sept. 2016 2016, pp. 1808-1814, doi:
10.1109/ICELMACH.2016.7732769.
[68] S. Dieltiens, J. D’hondt, M. Juwet, and M. Versteyhe,
"Development of a low-cost measurement system to
determine 3-dimensional pedal loads during in-situ cycling,"
Problemy Transportu, vol. 14, no. 4, pp. 151-159, 2019, doi:
10.20858/tp.2019.14.4.13.
[69] J. D'hondt, D. Degryse, P. Slaets, E. Demeester, and M. Juwet,
"Handling Qualities of a New Last-Mile Vehicle," Journal of
Transportation Technologies, vol. 12, no. 1, pp. 137-158,
2022, doi: 10.4236/jtts.2022.121009.
[70] J. P. Meijaard, J. M. Papadopoulos, A. Ruina, and A. L. Schwab,
"Linearized dynamics equations for the balance and steer of a
bicycle: a benchmark and review," Nelineinaya Dinamika, pp.
343-376, 2013, doi: 10.20537/nd1302010.
[71] S. Steinmaßl and M. Lienkamp, "Testing and Evaluation of a
New Multitrack Electric Bicycle—A Comparative Study," vol.
485, ed: Cham: Springer International Publishing, 2016, pp.
743-755.
[72] Cycles - Safety requirements for bicycles EN ISO 4210, E. C. f.
standardization, 2014.
[73] Passenger cars - Validation of vehicle dynamcs simulation -
Lateral transient response test methods, I. c. o. ISO22140,
2021.
97/107
[74] D. G. Wilson and T. Schmidt, Bicycling science, 4th edition ed.
Cambridge, Massachusetts : London, England : The MIT Press,
2020.
[75] J. Kooijman and A. Schwab, A Review on Handling Aspects in
Bicycle and Motorcycle Control. 2011.
[76] L. Li, X. Li, X. Wang, J. Song, K. He, and C. Li, "Analysis of
downshift’s improvement to energy efficiency of an electric
vehicle during regenerative braking," Applied Energy, vol. 176,
pp. 125-137, 2016/08/15/ 2016, doi:
https://doi.org/10.1016/j.apenergy.2016.05.042.
[77] T. U. O. T. S. O. N. YORK, "Reference Tables for Physical
Setting/PHYSICS," ed. Albany NY, 2006.
[78] Z. Xue, C. Li, X. Wang, L. Li, and Z. Zhong, "Coordinated control
of steer-by-wire and brake-by-wire for autonomous
emergency braking on split-μ roads," IET intelligent transport
systems, vol. 14, no. 14, pp. 2122-2132, 2020, doi:
10.1049/iet-its.2020.0184.
[79] V. E. Bulsink, A. Doria, D. van de Belt, and B. Koopman, "The
effect of tyre and rider properties on the stability of a bicycle,"
Advances in mechanical engineering, vol. 7, no. 12, p.
168781401562259, 2015, doi: 10.1177/1687814015622596.
[80] J. D'hondt, p. Slaets, e. Demeester, and m. Juwet, "Pedalling
Comfort of a Custom Pedal Series Hybrid Drivetrain in a Cargo
E-Tricycle," 2022.
[81] K. Augustus, "ELECTRICALLY POWERED CYCLE," U.S.A., 1975.
[82] A. Fuchs, "Electronic Bicycles are free of Costs for (automated)
mechanical Transmissions," in Bike Europe, ed. Berne,
Switzerland, 2012.
[83] S. Raghunath, "Hardware Design Considerations for an
Electric Bicycle Using a BLDC Motor," Texas Instruments,
2021. [Online]. Available:
https://www.ti.com/lit/an/slva642b/slva642b.pdf?ts=16521
23552078&ref_url=https%253A%252F%252Fwww.google.co
m%252F
[84] L. Cruces. "The arcimoto Lean Mean Machine." Arcimoto.
https://www.arcimoto.com/mlm (accessed May, 2022).
[85] R. dot, "E-Bike Mando Footloose," in Reddot design award:
best of the best 2013, ed. Red dot, 2013.
98/107
[86] E. b. review, "Mando Footloose Review," ed. Electric bike
review.
[87] Gondola, "Bubble Post deploys new cargo bicycle for city
distribution in Brussels," in Gondola vol. 2022, ed. Machelen:
Gondola, 2016.
[88] E. Tournon, "Optimal sizing of an hybrid series bicycle equiped
with supercapacitor," 2020.
[89] E. Tournon. "The solar bike of Edgar."
https://www.thesuntrip.com/en/participants/edgar-
tournon/ (accessed May, 2022).
[90] I. Lesens "un velo sans chaine ni batterie il pourrait
revolutionner la mobilite active."
https://www.isabelleetlevelo.fr/2021/01/05/un-velo-sans-
chaine-ni-batterie-il-pourrait-revolutionner-la-mobilite-
active/ (accessed February, 2022).
[91] B. Cauquil. "Mon Sun Trip 2021 en quelque chiffres."
https://www.thesuntrip.com/en/mon-sun-trip-2021-en-
quelques-chiffres/ (accessed January, 2022).
[92] B. Cauquil. "Sunrider EcoMobile Baroudeur."
http://www.sunrider.fr/ (accessed January, 2022).
[93] P. Brunet, "Le Teil : la ville essaie un vélobus électrique très
prisé en Europe," ed. Le Dauphiné, 2019.
[94] S. Cho and J. Chang. "Micro mobility." Mando.
https://www.mando.com/eng/rnd/rnd05.jsp (accessed
February, 2022).
[95] A. Schaeffler. "Schaeffler presents chainless electric drive
system “Free Drive” for bicycles." schaeffler.
https://www.schaeffler.com/en/news_media/press_release
s/press_releases_detail.jsp?id=87716736 (accessed
September, 2021).
[96] B. Borrell, "PHYSICS on TWO WHEELS," Nature (London), vol.
535, no. 7612, pp. 338-341, 2016, doi: 10.1038/535338a.
[97] A. L. Schwab and J. P. Meijaard, "A review on bicycle dynamics
and rider control," Vehicle system dynamics, vol. 51, no. 7, pp.
1059-1090, 2013, doi: 10.1080/00423114.2013.793365.
[98] T. Williams, S. Kaul, and A. Dhingra, "Influence of Frame
Stiffness and Rider Position on Bike Dynamics: Analytical
Study," presented at the ASME 2015 International Mechanical
99/107
Engineering Congress and Exposition, 2015. [Online].
Available: https://doi.org/10.1115/IMECE2015-50137.
[99] J. Xiong, N. Wang, and C. Liu, "Stability analysis for the
Whipple bicycle dynamics," Multibody System Dynamics, vol.
48, 03/01 2020, doi: 10.1007/s11044-019-09707-y.
[100] S. Dieltiens, J. D'hondt, M. Versteyhe, and M. Juwet,
"Theoretical self-stability of E-bikes," in 42ème Congrès de la
Société de Bioméchanique, Reims, France, 2017.
[101] J. D. G. Kooijman, J. P. Meijaard, J. M. Papadopoulos, A. Ruina,
and A. L. Schwab, "A Bicycle Can Be Self-Stable Without
Gyroscopic or Caster Effects," Science, vol. 332, no. 6027, pp.
339-342, 2011, doi: 10.1126/science.1201959.
[102] S. Dieltiens, J. D’hondt, M. Juwet, K. Shariatmadar, and M.
Versteyhe, "Design and calibration of an instrumented seat
post to measure sitting loads while cycling," Sensors (Basel),
vol. 20, no. 5, p. 1384, 2020, doi: 10.3390/s20051384.
[103] J. D'hondt, S. Dieltiens, m. Juwet, and M. Versteyhe, "Design
and Calibration of a 6-Component Balance on a Bicycle Steer,"
Proceedings of The Eighteenth International Conference of
Experimental Mechanics, vol. 2504-3900, pp. 520-520, 2018.
[104] D. Krumm and S. Odenwald, "Development of a
Dynamometer to Measure Grip Forces at a Bicycle
Handlebar," Procedia Engineering, vol. 72, pp. 80-85, 2014,
doi: 10.1016/j.proeng.2014.06.017.
[105] N. A. Turpin and B. Watier, "Cycling biomechanics and its
relationship to performance," Applied sciences, vol. 10, no.
12, pp. 1-15, 2020, doi: 10.3390/APP10124112.
[106] F. Höchtl, H. Böhm, and V. Senner, "Prediction of energy
efficient pedal forces in cycling using musculoskeletal
simulation models," Procedia Engineering, vol. 2, pp. 3211-
3215, 06/01 2010, doi: 10.1016/j.proeng.2010.04.134.
[107] S. Dorel, A. Couturier, J.-R. Lacour, H. Vandewalle, C. Hautier,
and F. Hug, "Force-Velocity Relationship in Cycling Revisited,"
Medicine and science in sports and exercise, vol. 42, pp. 1174-
83, 12/01 2009, doi: 10.1249/MSS.0b013e3181c91f35.
[108] R. Davis and M. Hull, "Measurement of pedal loading in
bicycling: II. Analysis and results," Journal of biomechanics,
vol. 14, pp. 857-72, 02/01 1981, doi: 10.1016/0021-
9290(81)90013-0.
100/107
[109] G. Mornieux, K. Zameziati, E. Mutter, R. Bonnefoy, and A.
Belli, "A cycle ergometer mounted on a standard force
platform for three-dimensional pedal forces measurement
during cycling," Journal of Biomechanics, vol. 39, no. 7, pp.
1296-1303, 2006/01/01/ 2006, doi:
https://doi.org/10.1016/j.jbiomech.2005.03.020.
[110] R. Bini, P. Hume, and A. Cerviri, "A comparison of cycling SRM
crank and strain gauge instrumented pedal measures of peak
torque, crank angle at peak torque and power output,"
Procedia Engineering, vol. 13, pp. 56-61, 12/31 2011, doi:
10.1016/j.proeng.2011.05.051.
[111] F. Bolourchi and M. Hull, "Measurement of Rider Induced
Loads during Simulated Bicycling," International Journal of
Sport Biomechanics, vol. 1, pp. 308-329, 11/01 1985, doi:
10.1123/ijsb.1.4.308.
[112] E. Favero. "Advantages of the IAV Power System."
https://cycling.favero.com/blog/tech-tests/whats-iav-power
(accessed February, 2022).
[113] E. Favero, "Influence of Angular Velocity of Pedaling on the
Accuracy of the Measurement of Cyclist Power," 2021.
[114] S. Dieltiens, C. Jiménez-Peña, S. Van Loon, J. D’hondt, K.
Claeys, and E. Demeester, "Influence of Electrically Powered
Pedal Assistance on User-Induced Cycling Loads and Muscle
Activity during Cycling," Applied Sciences, vol. 11, no. 5, 2021,
doi: 10.3390/app11052032.
[115] J. D'hondt, p. Slaets, E. Demeester, and M. Juwet, "Effects of
a Torsion Spring Used in a Flexible Delta Tricycle," 2022.
[116] V. Naumov, "Substantiation of Loading Hub Location for
Electric Cargo Bikes Servicing City Areas with Restricted
Traffic," Energies, vol. 14, no. 4, 2021, doi:
10.3390/en14040839.
[117] J. Ferguson. "Cargo Bike Crazy: The Potential of Delivering
Goods By Bike." https://ecf.com/news-and-
events/news/cargo-bike-crazy-potential-delivering-goods-
bike-0 (accessed March, 2022).
[118] M. A. Rodríguez Licea, E. A. Vazquez Rodríguez, F. J. Perez
Pinal, and J. Prado Olivares, "The Rollover Risk in Delta
Tricycles: A New Rollover Index and Its Robust Mitigation by
Rear Differential Braking," Mathematical Problems in
101/107
Engineering, vol. 2018, p. 4972419, 2018/10/17 2018, doi:
10.1155/2018/4972419.
[119] J. Oortwijn, "Development Starts for Cargo Bikes' Safety
Standard," in Bike-EU, ed. Delft: Bike Europe, 2017.
[120] Sevic S180 catalogue, 2019.
[121] Volkswagen, "Local logistics made easy: the Volkswagen
Cargo e-bike," ed, 2018.
[122] I. Bae, J. Moon, and J. Seo, "Toward a Comfortable Driving
Experience for a Self-Driving Shuttle Bus," Electronics, vol. 8,
no. 9, 2019, doi: 10.3390/electronics8090943.
[123] S. M. Cain, J. A. Ashton-Miller, and N. C. Perkins, "On the skill
of balancing while riding a bicycle," PLoS One, vol. 11, no. 2,
pp. e0149340-e0149340, 2016, doi:
10.1371/journal.pone.0149340.
[124] S. R. Vejgaard, B. Basalisco, A. M. Boivie, and H. B. Okholm, "E-
commerce and delivery," Copenhagen Economics, 2013.
[Online]. Available:
https://copenhageneconomics.com/publication/e-
commerce-and-delivery/
[125] B. Times. "The rise of the cargo bike: Companies rethink inner-
city deliveries." The Brussels Time.
https://www.brusselstimes.com/226021/the-rise-of-the-
cargo-bike-companies-rethink-inner-city-deliveries (accessed
August, 2022).
[126] A. Budholiya. "Accounting for 43% of Sales, Courier and Parcel
Service Providers Set to Dominate Global Cargo Bike Market
as Leading End Users: Future Market Insights." PRNewswire.
https://www.prnewswire.com/news-releases/accounting-
for-43-of-sales-courier-and-parcel-service-providers-set-to-
dominate-global-cargo-bike-market-as-leading-end-users-
future-market-insights-301353316.html (accessed August,
2022).
[127] M. Placek. "Total volume of parcel and express services in
Belgium from 2010 to 2020."
https://www.statista.com/statistics/752220/volume-of-
parcel-and-express-services-in-belgium/ (accessed August,
2022).
[128] P. Hung. "E-Commerce Trends 2022: What The Future Holds."
Forbes.
102/107
https://www.forbes.com/sites/forbestechcouncil/2022/03/1
4/e-commerce-trends-2022-what-the-future-holds/
(accessed August, 2022).
[129] J. Gruber, "Ergebnisse des größten Lastenradtests
Europas Umweltfreundlich und effizient: Lastenräder im
dienstlichen und gewerblichen Einsatz," ed, 2021.
[130] M. Sheth, P. Butrina, A. Goodchild, and E. McCormack,
"Measuring delivery route cost trade-offs between electric-
assist cargo bicycles and delivery trucks in dense urban areas,"
European Transport Research Review, vol. 11, no. 1, p. 11,
2019/02/13 2019, doi: 10.1186/s12544-019-0349-5.
[131] D. Carter, "Massive jump in online shopping as Belgian market
grows by 33% in 2021," in The Brussels Times, ed. Brussels,
2022.
[132] I. Mordor, "Belgium e-commerce market - growth, trends,
covid-19 impact, and forecasts (2023 - 2028)," ed: Mordor
Intelligence, 2022.
[133] A.-K. Reibold, "E-Commerce Delivery Emissions Could
Increase by 30% Without Intervention."
https://ecf.com/news-and-events/news/e-commerce-
delivery-emissions-could-increase-30-without-
intervention
[134] D. Nowak. "How sustainable is the Pedal Me fleet?"
https://pedalme.co.uk/carbon-emissions/ (accessed August,
2022).
[135] L. V. Harry. "Bullitt cargo bikes deliver faster, cheaper and
cleaner than vans, study finds."
https://www.larryvsharry.com/da/blog/last-mile-delivery-
study (accessed March, 2024).
[136] D. Carrington, "Cargo bikes deliver faster and cleaner than
vans, study finds," in The guardian, ed, 2021.
[137] R. Lowe and M. Rigby. "The Last Mile Exploring the online
purchasing and delivery journey." Barclays.
https://www.utikad.org.tr/images/BilgiBankasi/thelastmilee
xploringtheonlinepurchasinganddeliveryjourney-7912.pdf
(accessed March, 2023).
[138] G. Charlton and B. Ward. "Which Ecommerce Online Delivery
Options Do Customers Want?" SaleCycle.
https://www.salecycle.com/blog/strategies/ecommerce-
delivery-options-customers-want/ (accessed April, 2023).
103/107
[139] A. Behrensen, "3rd European Cargo Bike Industry Survey,"
Eurobike, Frankfurt, 2022. [Online]. Available:
http://cyclelogistics.eu/market-size
[140] G. V. Research, "Bicycle Market Size, Share & Trends Analysis
Report By Product (Mountain, Hybrid, Road, Cargo), By
Technology (Electric, Conventional), By End User, By
Distribution Channel, By Region, And Segment Forecasts,
2022 - 2030," ed, 2022.
[141] NieuwsFiets.nu. "Share of bicycle sales in Belgium in 2019, by
type." Statista.
https://www.statista.com/statistics/620542/share-of-
bicycle-sales-in-belgium-by-type/ (accessed July, 2022).
[142] P. M. Research. "Electric cargo bike market."
https://www.persistencemarketresearch.com/market-
research/electric-cargo-bikes-market.asp (accessed July,
2022).
[143] N. kaitwade. "Cargo bike market." Future market insights.
https://www.futuremarketinsights.com/reports/cargo-bike-
market (accessed February, 2023).
[144] T. Brainy Insights. "cargo bike market." GlobeNewswire.
https://www.thebrainyinsights.com/report/cargo-bike-
market-
12794#:~:text=The%20global%20cargo%20bike%20market,1
3.7%25%20from%202022%20to%202030. (accessed August,
2022).
[145] V. Overheid. "Omzet via e-commerce bij ondernemingen."
Vlaanderen. https://www.vlaanderen.be/statistiek-
vlaanderen/digitale-economie/omzet-via-e-commerce-bij-
ondernemingen (accessed March, 2023).
[146] F. Rylant, K. De Belder, S. Lampaert, and L. Bisschop, "De
belgische fietsmarkt in 2022," Traxio, 2023.
[147] No 168/2013, T. E. PARLIAMENT and COUNCIL, Official Journal
of the European Union, 2013.
[148] A. Roetynck, "Electric Cargo Bikes: Technical Rules &
Regulations," in ECLF conference, Vienna, 2017, 2017 ed.
[Online]. Available:
https://eclf.bike/presentations17/D1%20Annick%20Roetync
k%20ECLF%202017.pdf
105/107
Figures
Figure 1 Bike vs. car shipments: (left) cumulative frequencies of single
shipment distances (right) cumulative frequencies of aggregated
daily mileages [23]. ....................................................................... 7
Figure 2 What was the approximate weight of this particular purchase [29]?
....................................................................................................... 8
Figure 3 Cargo bike types [13]. ...................................................................... 9
Figure 4 Representation of cargo bike types [49]. ........................................ 10
Figure 5 Ziegler Cargo Bike XXL [57] and Urban Arrow Tender [58]. ...... 11
Figure 6 Concept of the new type of cargo bike. .......................................... 14
Figure 7 Handlebar of the tricycle. ............................................................... 18
Figure 8 Prototype with adjustable hinge. .................................................... 21
Figure 9 Turning radius of the vehicle with a) a downwards tilted hinge, b) a
horizontal hinge and c) an upwards tilted hinge. ......................... 22
Figure 10 Last-Mile Vehicle prototype. ....................................................... 23
Figure 11 LMV concept. .............................................................................. 25
Figure 12 LMV prototype. ............................................................................ 27
Figure 13 QuadRad, City Shopping Bike, Babboe Big, Babboe City (from left
to right) [71]. ................................................................................ 27
Figure 14 Legend of manoeuvre setup. ........................................................ 29
Figure 15 Emergency brake. ......................................................................... 30
Figure 16 µ-split brake. ................................................................................ 32
Figure 17 Steady-state turn. .......................................................................... 33
Figure 18 Figure of eight. ............................................................................. 34
Figure 19 Prompt steering. ........................................................................... 34
Figure 20 Double lane change. ..................................................................... 35
Figure 21 Slalom. ......................................................................................... 36
Figure 22 Motorcycle Manoeuvring circuit. ................................................. 37
Figure 23 µ-split brake hub motor speed. ..................................................... 39
Figure 24 Lateral acceleration of City shopping during steady-state turn [71].
..................................................................................................... 40
Figure 25 Lateral acceleration of Babboe Big during steady-state turn [71].
..................................................................................................... 41
Figure 26 Lateral acceleration of QuadRad during steady-state turn [71]. ... 41
Figure 27 Lateral acceleration of LMV driver and Cargo during steady-state
turn. .............................................................................................. 41
Figure 28 Lateral acceleration of LMV driver and cargo during figure of eight.
..................................................................................................... 42
Figure 29 Pedal Series Hybrid workings [64]............................................... 48
Figure 30 Mando footloose [85]. .................................................................. 50
Figure 31 Tortuga XL [87]. .......................................................................... 50
Figure 32 a) Vehicle of Edgar Tournon [89] b) Bernard Cauquil [92]. ........ 51
106/107
Figure 33 Bicycle bus [93]. .......................................................................... 51
Figure 34 a) Mando SPM [94] b) Schaeffler free drive [95]. ....................... 52
Figure 35 Set of feasible forces at different crank angles [105]. .................. 53
Figure 36 Right pedal forces and torque profiles (a) Fz (b) Fy (c) Fx (d) torque
with 0° being the upward pedal position [111]. ........................... 53
Figure 37 Crank torque measurement by the pedals and SRM torque analysis
[110]. ........................................................................................... 54
Figure 38 a) Force and angular velocity of IAV power meter [112] b) angular
velocity and tangential force of a single pedal revolution of 5
different cyclists [113]. ................................................................ 55
Figure 39 Delta tricycle cargo vehicle. ......................................................... 56
Figure 40 Pedal generator control loop. ........................................................ 57
Figure 41 Wheel control loop. ...................................................................... 58
Figure 42 Force/current according to orientation and T-Test results. ........... 60
Figure 43 Graph of angular velocity variations and T-Test results. ............. 61
Figure 44 Last-mile vehicle concept [69]. .................................................... 64
Figure 45 Lateral inclination, downhill inclination and uphill inclination. .. 65
Figure 46 Torsion axle and corresponding spring characteristic. ................. 68
Figure 47 Tilting concept vehicle. ................................................................ 69
Figure 48 Measured torque/angle curve with corresponding upper trendline
t=0.0013·a³+0.016·a²+0.0851·a+2.3532. with t the torsion (Nm)
and a the torsion angle (°). ........................................................... 70
Figure 49 Tilting driver vehicle while parked. ............................................. 71
Figure 50 Stationary tilting curve on a level and inclined surface. ............... 72
Figure 51 Tilting angle according to turning radius and velocity. Calculations
are performed at 5, 10, 15, 20, and 25 km/h. ............................... 74
Figure 52 Graph of the current and 10 times stiffer torsion system. These are
compared with the torsion the driver module with and without
cyclist creates on level surface and inclined (8%). ...................... 76
Figure 53 Tilting calculation of torsion system and 10 times stiffer torsion
system and difference due to torsion system. .............................. 76
107/107
Tables
Table 1 Technical specifications .................................................................. 28
Table 2 Braking requirement ........................................................................ 31
Table 3 Summary of manoeuvres ................................................................. 36
Table 4 Full rear brake application 25 km/h ................................................. 38
Table 5 Steady state-turn .............................................................................. 40
Table 6 Prompt steering ................................................................................ 43
Table 7 Double lane change ......................................................................... 43
FACULTY OF ENGINEERING TECHNOLOGY
DEPARTMENT OF MECHANICAL ENGINEERING
INNOVATIVE TECHNOLOGY FOR LOGISTICS
Gebroeders de Smetstraat 1
9000 Ghent, BELGIUM
ordi.dhondt@kuleuven.be
https://iiw.kuleuven.be/onderzoek/itl
