Content uploaded by Priyanshu Sinha
Author content
All content in this area was uploaded by Priyanshu Sinha on Mar 25, 2025
Content may be subject to copyright.
Content uploaded by Siddharth Garg
Author content
All content in this area was uploaded by Siddharth Garg on Mar 18, 2025
Content may be subject to copyright.
Mechanical Engineering for Sustainable Development: ISME-2024 Materials Research Forum LLC
Materials Research Proceedings 49 (2025) 350-359 https://doi.org/10.21741/9781644903438-35
Content from this work may be used under the terms of the Creative Commons Attribution 3.0 license. Any further distribution of
this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. Published under license by Materials
Research Forum LLC.
350
Design and development of drivetrain for fully faired hybrid
recumbent bicycle
Siddharth GARG1,a*, Rhythm AGGARWAL1,b, Priyanshu SINHA2,c,
Vaibhav KUMAR1,d, Raghvendra GAUTAM1,e
1Department of Mechanical Engineering, Delhi Technological University, Delhi 110042, India
2Department of Electrical Engineering, Delhi Technological University, Delhi 110042, India
aofficialsiddharthgarg@gmail.com, brhythm4901@gmail.com, cpriyanshusinha.dtu@gmail.com,
dvaibhav.ky.2003@gmail.com, eraghvendrag80@yahoo.com
Keywords: Recumbent Bicycle, Velomobile, Hybrid Drivetrain, MATLAB, HEV
Abstract. Due to the rapid consumption of fossil fuels, the effects of climate change have become
very rampant in recent years. Harvesting human and electric energy in the form of hybrid electric
vehicles (HEVs) is a cleaner and more economical solution to the problem of urban mobility. Fully
faired recumbent bicycles, also called velomobiles, while retaining all the good features of bicycles
(eco-friendly, small frontal area), also incorporate the beneficial features of cars (protection from
weather and crashes, extra added stability). This study examines the three drivetrains possible for
velomobiles (fully electric, fully human-powered, and hybrid, the latter of which incorporates both
human and electric energy). It compares them on the parameters of maximum speed, acceleration,
range, and driving power needed. The simulations were designed in MATLAB Simulink, and the
results were verified by actual human testing performed under suitable conditions. The testing
revealed that the hybrid drivetrain model is the most efficient to use and to satisfy all the desired
benchmarks of ASME e-HPVC.
INTRODUCTION
Amidst a rapidly escalating climate crisis, the world stands at a critical juncture. Central to this
environmental upheaval is the pervasive use of fossil fuels in vehicles, expelling greenhouse gases
at an unprecedented rate. Therefore, governments across the globe have been pushing for more
sustainable modes of transportation. One of the areas of research has been biofuels [1–4]; although
they have shown promising results, none of them has been able to match hybrid electric vehicles
(HEVs) [5]. The main objective of this study is to examine possible drivetrains for fully-faired
recumbent bicycles. Recumbents have been in use for an extended period but have not seen any
upgrade from their original design, which is one of the reasons for their loss in popularity [6]. The
incorporation of a hybrid drivetrain will make it a suitable contender for urban mobility, where the
distance to be travelled is small, and only one passenger space is required [7]. In this work, an
electric human-powered fully faired recumbent is designed in the context of participating in the
American Society of Mechanical Engineers Electric-Human Powered Vehicle Challenge (ASME
e-HPVC). This study outlines the development process of three Simulink models for drivetrains,
which were designed in MATLAB. They are tested on various criteria, including range,
acceleration, speed and driving power. Based on the results, an actual model for vehicle testing
(Fig. 1) is constructed. The simulation results are verified experimentally by members of Team
Raftaar of Delhi Technological University. The main contribution of this study is the in-depth
description of the design for an efficient hybrid drivetrain for recumbent bicycles using a model-
based design methodology and software modelling and simulation, along with verification of the
results via actual testing.
Mechanical Engineering for Sustainable Development: ISME-2024 Materials Research Forum LLC
Materials Research Proceedings 49 (2025) 350-359 https://doi.org/10.21741/9781644903438-35
351
Fig. 1. 3D CAD model of the fully-faired hybrid recumbent bicycle used for actual testing
purposes.
EXPERIMENTAL SETUP
Simulink Modelling
MATLAB Simulink is a graphical programming environment used for modelling, simulating, and
analysing dynamic systems and control algorithms. Three possible drivetrains are modelled in the
software as mentioned below in Table 1.
Table 1: Input values for the Simulink model
Parameter
Value
Weight
110 kg (70 kg + 40 kg)
Coefficient of Drag
0.063
Coefficient of Rolling Resistance
0.04
Density of air
1.127 kg/m3
Frontal Area
0.473 m2
Input torque
variable
Aux load
0
Efficiency
1
Gear Ratio (for human powered and
hybrid)
4
Radius of Tyre
0.66 m
(a) Fully Human Powered
In the Simulink model for fully human powered vehicle (Fig. 2), the required torque is taken as an
input in the form of pedalling and converted into force (Eqs. 1 and 2) [8]. Aerodynamic and rolling
resistance forces are then subtracted to determine the net force acting on the wheels. The net force
is integrated to obtain the acceleration, which is further integrated to determine the velocity. The
power is calculated by multiplying the net force and velocity. The change in distance travelled is
used to calculate the crank rotation, allowing the model to determine the force applied by the rider.
This force is then converted back to torque, and the cycle continues.
Human Input Torque = Length of Crank (Lc) × Chain Force × Gear Ratio (1)
Mechanical Engineering for Sustainable Development: ISME-2024 Materials Research Forum LLC
Materials Research Proceedings 49 (2025) 350-359 https://doi.org/10.21741/9781644903438-35
352
Chain Force = ½ × (cos 2(Crank Angle) + 1) × Rider’s Weight (2)
Fig. 2. Simulink model for a human powered drivetrain for the recumbent bicycle
(b) Fully Electric Powered
In this model (Fig. 3), the motor managed by a throttle provides the necessary torque (Eq. 3). To
determine the values of acceleration, velocity and acceleration, the net force that remains after
removing the frictional and drag losses is integrated multiple times and the values at different
stages are noted (Eq. 4) [9]. Consequently, we calculated the required driving power by
multiplying net effective force and vehicle speed (Eq.5) [10]. By operating it with the battery
capacity, we arrived at the vehicle range (Eq.6).
Electric Input Torque = Input Force × Radius of Tyre (3)
Net Force (F) = Input Force – (Aerodynamic Drag + Rolling Resistance) (4)
Driving Power = Input Force × Velocity (V) (5)
Range = ×
( + ) × (6)
Mechanical Engineering for Sustainable Development: ISME-2024 Materials Research Forum LLC
Materials Research Proceedings 49 (2025) 350-359 https://doi.org/10.21741/9781644903438-35
353
Fig. 3. Simulink model for an electric powered drivetrain for a recumbent bicycle
(c) Hybrid Drivetrain ( Human and Electric Powered)
Fig. 4 presents the hybrid model proposed here, which amalgamates human-powered and motor-
assisted aspects for comprehensive vehicle dynamics analysis. The system takes input torque from
pedalling and transforms it into force. The net force acting on the wheels is derived by accounting
for the aerodynamic and rolling resistance forces. Integration techniques are employed to ascertain
acceleration and velocity, which, in turn, lead to power computation (Eq. 7) [11].
Hybrid Input Torque = Human Input Torque + Electric Input Torque (7)
Fig. 4. Simulink model for the hybrid drivetrain (both human and electric power are used in
combination) for the fully faired recumbent bicycle
Vehicle Fabrication
The vehicle under study is a 2-wheeler, rear-wheel drive recumbent bicycle, as depicted in Fig. 5.
The crank set is placed in front of the vehicle. The frame of the vehicle had a wheelbase of 1.474
m and a trail of 0.05 m. The gear ratio for the sprocket and the crank is 4:1. A 48 V 500 W hub
motor was used in the hub of the rear tyre. The motor was controlled by a 48V~35A motor
controller and 48V 10Ah Battery. Air-dampened, short travel suspension on the rear tyre was used
Mechanical Engineering for Sustainable Development: ISME-2024 Materials Research Forum LLC
Materials Research Proceedings 49 (2025) 350-359 https://doi.org/10.21741/9781644903438-35
354
for all-round performance [12–15]. The vehicle used 26-inch (0.66 m) rear and 20-inch (0.508 m)
diameter front tyres (inflation pressure of 30 PSI). All subjects utilised the same chassis during the
experiments (a carbon fibre and titanium frame of 1.8 kg). Disc brakes of 180 mm on both wheels
were used. Over the seat, direct steering mechanism with a steering angle of 72° was used for
precise control and to reduce weight [5,16,17]. The chassis was fully faired with a 40 kg, 12-piece
carbon fibre fairing, specially manufactured using the vacuum bagging technique, additive
manufacturing and brazing for custom joints [18–21].
Fig. 5. Picture depicting detailed schematic of the drivetrain of recumbent bicycle
Testing Conditions
The rider was 20 years of age, had a 70 kg body mass, and was 173 cm tall. The rider was part of
the research and development team and clearly understood the risks involved. Testing was carried
out according to the standards set for ASME for the e-HPVC. The experiments were performed on
a road with 3 km of straight stretch with no elevation. The wind speed was recorded to be 18 kmph.
A Garmin cycle speedometer edge 1030 was used to measure the speed. The testing began every
day at approximately 10:00 hours and were concluded at about 15:00 hours [22–24]. An eight
samarium-cobalt magnet PAS sensor was mounted on the chainring to measure the intensity of the
pedals. The motor controller then uses this information to supply the motor with the necessary,
proportional current.
Fig. 6. Actual testing of vehicle carried out at Delhi Technological University campus by the
members of Team Raftaar
Mechanical Engineering for Sustainable Development: ISME-2024 Materials Research Forum LLC
Materials Research Proceedings 49 (2025) 350-359 https://doi.org/10.21741/9781644903438-35
355
RESULTS
From the above simulation, various parameters and graphs for comparison were obtained, which
were then utilised while implementing our actual recumbent vehicle.
Fig. 7. Graph between the driving power and velocity for all three cases.
Fig. 7 depicts the required driving power against the vehicle's velocity graph plotted for the
three powertrains: human-powered, electric, and hybrid. The graph clearly illustrates that the
driving power is directly proportional to the vehicle's velocity in all three cases. Notably, the hybrid
model demonstrates the highest velocity compared to the human-powered and electric models.
This finding confirms that the hybrid powertrain, which combines human and electric propulsion,
achieves the maximum velocity among the three configurations.
Fig. 8. Graph between time and acceleration for all three cases
The analysis of the time versus acceleration (Fig. 8) and time versus velocity (Fig. 9) graphs
reveals a common trend across all models, where they exhibit an initial period where both
acceleration and velocity experience changes before stabilizing. During the initial period, vehicle
takes some time to gain velocity from rest, and changes in acceleration accompany this process.
The acceleration is not constant during this transient phase but shows fluctuations as the vehicle
accelerates and overcomes resistive forces such as friction and air drag. However, after this initial
Mechanical Engineering for Sustainable Development: ISME-2024 Materials Research Forum LLC
Materials Research Proceedings 49 (2025) 350-359 https://doi.org/10.21741/9781644903438-35
356
transient period, both the acceleration and velocity reach a steady state, where they become
constant. In this state, the vehicle maintains a constant acceleration, corresponding to a consistent
velocity increase over time.
Fig. 9. Graph between the RPM and Net input torque for all three case
In Fig. 9, graph's net input torque represents the value obtained after subtracting the drag and
resistance forces from the required forces. The graph exhibits an interesting pattern, where
initially, at the starting rotations per minute (RPM) (which is zero), the net torque is at its maximum
value, equivalent to the required torque for the given driving conditions. As the speed of the vehicle
increases, the drag and resistance forces also increase, resulting in a decrease in the net torque. As
the vehicle continues to accelerate, the net torque gradually approaches zero. Simultaneously, the
RPM of the vehicle steadily increases as it gains speed. The graph shows how the net torque
decreases as the RPM increases, indicating that the vehicle requires less torque to maintain higher
RPM levels due to reduced resistive forces at higher speeds. Consistent with the previous findings,
the hybrid model again has the highest RPM among the three powertrains.
The hybrid model was implemented in the vehicle based on the comprehensive analysis of the
various graphs and parameters obtained from the simulations. It has demonstrated numerous
advantages over human-powered and electric models regarding range, velocity, and maximum
RPM. The experimental results of the actual testing data and simulated data are tabulated in Table
2.
Table 2: Comparison of the results of all the drivetrain models
Properties Human Powered
(Simulink)
Electric
(Simulink)
Hybrid
(Simulink)
Hybrid
(Experimental)
Required Torque (Nm) 3.04 9.5 12.54 (9.5 + 3.04) 10.2
Driving Power (Watt) 39 350 559 524
Max. Velocity (km/hr) 45 88 106 97
RPM
110.5
352.72
426.42
389.84
Range (km)
-
53.29
70.15
75.25
Mechanical Engineering for Sustainable Development: ISME-2024 Materials Research Forum LLC
Materials Research Proceedings 49 (2025) 350-359 https://doi.org/10.21741/9781644903438-35
357
For the results tabulated in Table 2, the required torques were set to 3.04 Nm, 9.5 Nm, and
12.54 Nm for the human-powered, electric, and hybrid Simulink models, respectively. The
experimental error for all the results was under 10% (arising due to varying air resistance and
friction), indicating the efficacy of the model. When analysing the hybrid model, a combination of
torques from human-powered and electric models was utilised. As a result, there were improved
outcomes compared to those of the electric model, specifically regarding range and velocity.
Remarkably, this improvement is achieved while maintaining the same input torque and driving
power from the electric motor.
CONCLUSIONS
This study systematically compared three Simulink models representing electric, human-powered,
and hybrid bicycles while supplementing the analysis with a real-life experiment on the hybrid
variant. The findings unequivocally demonstrate that the hybrid bicycle outperforms both electric
and human-powered counterparts in key technical aspects such as range, velocity, and power. The
simulations provided valuable insights into the theoretical performance of each drivetrain type,
revealing the advantages and limitations of each. However, the real-life experiment on the hybrid
bicycle provided empirical evidence with an error percentage metric of 8.66%, 6.26%, 8.49%,
8.58%, and 7.27% for torque, power, velocity, RPM, and range, respectively, validating the
superiority of the simulations and indicating the high accuracy and reliability of the findings.
Considering the results from both the simulations and the real-life experiment, the hybrid bicycle
emerged as the most promising option for various applications, especially when a balance between
range, velocity, and power is desired. The ability of HEVs to leverage both electric power and
human input showcases a sustainable and efficient means of transportation.
ACKNOWLEDGMENT
This work was supported by Delhi Technological University, Delhi-110042, India (sanction order
number DTU/DSW/2022-23/F.No.90/132)
DECLERATION
Competing Interest: The authors declare no competing interests.
Author Contribution: All authors contributed to the study conceptualisation and vehicle
fabrication. S.Garg: Manuscript writing and editing, vehicle testing. R.Aggarwal: Manuscript
writing and MATLAB simulation P.Sinha: MATLAB simulation V.Kumar: Vehicle testing and
CAD modelling R.Gautam: Manuscript editing and supervision. All authors read and approved
the final manuscript.
REFERENCES
[1] R. Gautam, N.A. Ansari, P. Thakur, A. Sharma, Y. Singh, Status of biofuel in India with
production and performance characteristics: a review, International Journal of Ambient Energy
43 (2022) 61–77. https://doi.org/10.1080/01430750.2019.1630298
[2] D. Sikha, R. Gautam, S. Kumar, Performance And Combustion Analysis Of Diesel And
Sesame Biodiesel In Small Capacity Diesel Engine, Design Engineering (2021) 4923–4938.
[3] R. Gautam, S. Kumar, Performance and combustion analysis of diesel and tallow biodiesel
in CI engine, Energy Reports 6 (2020) 2785–2793.
https://doi.org/https://doi.org/10.1016/j.egyr.2020.09.039
[4] S. Kumar, R. Gautam, Prospects of Factor Affecting Biodiesel Selection Strategies Based on
Various Aspects: An Indian Perspective, Journal of Engineering Research (Kuwait) 10 (2022)
208–219. https://doi.org/10.36909/jer.ICAPIE.15043
Mechanical Engineering for Sustainable Development: ISME-2024 Materials Research Forum LLC
Materials Research Proceedings 49 (2025) 350-359 https://doi.org/10.21741/9781644903438-35
358
[5] Y. Sharma, R. Yadav, S. Verma, M. Sehgal, R. Gautam, Study About the Future of Electric
Vehicles in the Current Indian Scenario, SAE Technical Paper (2019).
[6] H. Ahmed, O.M. Qureshi, A.A. Khan, Reviving a ghost in the history of technology: The
social construction of the recumbent bicycle, Soc Stud Sci 45 (2015) 130–136.
https://doi.org/10.1177/0306312714560640
[7] K. Singh, N.S. Iyengar, D. Ashok, Design and Analysis of Human Powered Hybrid Vehicle,
International Journal of Mechanical Engineering and Technology (IJMET 9 (2018) 594–605.
https://doi.org/10.13140/RG.2.2.29245.36320
[8] P.D. Soden, B.A. Adeyefa, Forces applied to a bicycle during normal cycling, J Biomech 12
(1979) 527–541. https://doi.org/10.1016/0021-9290(79)90041-1
[9] J.B. Spicer, C.J.K. Richardson, M.J. Ehrlich, J.R. Bernstein, M. Fukuda, M. Terada, Effects
of frictional loss on bicycle chain drive efficiency, Journal of Mechanical Design, Transactions
of the ASME 123 (2001) 598–605. https://doi.org/10.1115/1.1412848
[10] F. Valero, F. Rubio, C. Llopis-Albert, J.I. Cuadrado, Influence of the Friction Coefficient on
the Trajectory Performance for a Car-Like Robot, Math Probl Eng 2017 (2017).
https://doi.org/10.1155/2017/4562647
[11] V. Sankaranarayanan, S. Ravichandran, Torque sensorless control of a human-electric
hybrid bicycle, in: 2015 International Conference on Industrial Instrumentation and Control
(ICIC), IEEE, 2015: pp. 806–810. https://doi.org/10.1109/IIC.2015.7150852
[12] B. Howey, B. Bilgin, A. Emadi, Design of an External-Rotor Direct Drive E-Bike Switched
Reluctance Motor, IEEE Trans Veh Technol 69 (2020) 2552–2562.
https://doi.org/10.1109/TVT.2020.2965943
[13] B.C. Langford, C.R. Cherry, D.R. Bassett, E.C. Fitzhugh, N. Dhakal, Comparing physical
activity of pedal-assist electric bikes with walking and conventional bicycles, J Transp Health 6
(2017) 463–473. https://doi.org/10.1016/j.jth.2017.06.002
[14] J. Partyka, Comparative Analysis of the Mobility Assessment Methods for Tracked
Vehicles, in: Materials Research Proceedings, Association of American Publishers, 2022: pp.
221–226. https://doi.org/10.21741/9781644902059-32
[15] E. Tournon, P. Venet, B. Barbedette, A. Lelievre, J. Aubry, A. SARI, Efficiency
Comparison between Series Hybrid Bike and Traditional Bike, in: 2019 Fourteenth International
Conference on Ecological Vehicles and Renewable Energies (EVER), IEEE, 2019: pp. 1–7.
https://doi.org/10.1109/EVER.2019.8813618
[16] M.S.M. Sani, N.A. Nazri, S.N. Zahari, N.A.Z. Abdullah, G. Priyandoko, Dynamic Study of
Bicycle Frame Structure, in: IOP Conf Ser Mater Sci Eng, Institute of Physics Publishing, 2016.
https://doi.org/10.1088/1757-899X/160/1/012009
[17] N.R. Miller, D. Ross, The Design of Variable-Ratio Chain Drives for Bicycles and
Ergometers—Application to a Maximum Power Bicycle Drive, Journal of Mechanical Design
102 (1980) 711–717. https://doi.org/10.1115/1.3254810
[18] S. Garg, A. Sardar, R. Srivastava, S. Verma, A.K. Madan, Variation in Tensile Strength of
3D Printed PLA Parts by Varying Infill Density and Infill Pattern, Saudi Journal of Engineering
and Technology 8 (2023) 103–107. https://doi.org/10.36348/sjet.2023.v08i05.004
Mechanical Engineering for Sustainable Development: ISME-2024 Materials Research Forum LLC
Materials Research Proceedings 49 (2025) 350-359 https://doi.org/10.21741/9781644903438-35
359
[19] S. Garg, Q. Murtaza, Comparative Analysis of Dip-Brazing and TIG Welding on the
Properties of Al-64430 Joints, J Adhes Sci Technol (2024) 1–15.
https://doi.org/10.1080/01694243.2024.2409317
[20] S. Garg, Q. Murtaza, Effect of filler paste’s mixing ratio on the properties of Al-64430 dip-
brazed joints, Welding in the World 68 (2024) 2459–2471. https://doi.org/10.1007/s40194-024-
01772-y
[21] S. Garg, S. Bansal, Q. Murtaza, Failure investigation of an elbow pipe used in sewage water
treatment facility, Materials and Corrosion 75 (2024) 1185–1192.
https://doi.org/10.1002/maco.202414336
[22] M.-K. Tran, M. Akinsanya, S. Panchal, R. Fraser, M. Fowler, Design of a Hybrid Electric
Vehicle Powertrain for Performance Optimization Considering Various Powertrain Components
and Configurations, Vehicles 3 (2020) 20–32. https://doi.org/10.3390/vehicles3010002
[23] C. Capelli, L.P. Ardigò, F. Schena, P. Zamparo, Energy cost and mechanical efficiency of
riding a human-powered recumbent bicycle, Ergonomics 51 (2008) 1565–1575.
https://doi.org/10.1080/00140130802238614
[24] E. Bressel, S. Bliss, J. Cronin, A field-based approach for examining bicycle seat design
effects on seat pressure and perceived stability, Appl Ergon 40 (2009) 472–476.
https://doi.org/10.1016/j.apergo.2008.10.001
