PreprintPDF Available
Preprints and early-stage research may not have been peer reviewed yet.

Abstract and Figures

In order to investigate the interaction between automated vehicles (AVs) and bicyclists, we present a coupled driving simulator that enables these two traffic participants to interact in a virtual environment. To avoid potentially dangerous situations in road traffic, human perception can be extended by communication between vehicles and their environment. In order to assist the communication process between traffic participants, mobile devices are applied as human-machine interfaces (HMIs). The simulator links the simulation and visualization software with a web application to control the HMIs. The passenger of the AV can change priority rules at conflict situation in the simulation with that application and therefore influence the vehicles behavior via the communication application. To test the coupled simulator, a proof of concept study with 16 simulation runs and two participants each is conducted. The subjects rated the overall simulation impressions tending positive. Based on the evaluation of the study participants, the simulator setup will be further developed.
Content may be subject to copyright.
A coupled driving simulator to investigate the interaction between
bicycles and automated vehicles*
Johannes Lindner1, Andreas Keler1, Georgios Grigoropoulos1, Patrick Malcolm1,
Florian Denk2, Pascal Brunner2, Klaus Bogenberger1
Abstract In order to investigate the interaction between
automated vehicles (AVs) and bicyclists, we present a coupled
driving simulator that enables these two traffic participants
to interact in a virtual environment. To avoid potentially
dangerous situations in road traffic, human perception can
be extended by communication between vehicles and their
environment. In order to assist the communication process
between traffic participants, mobile devices are applied as
human-machine interfaces (HMIs). The simulator links the
simulation and visualization software with a web application
to control the HMIs. The passenger of the AV can change
priority rules at conflict situation in the simulation with that
application and therefore influence the vehicles behavior via
the communication application. To test the coupled simulator,
a proof of concept study with 16 simulation runs and two
participants each is conducted. The subjects rated the overall
simulation impressions tending positive. Based on the evaluation
of the study participants, the simulator setup will be further
developed.
I. INTRODUCTION
Driving simulators are a valuable tool in traffic research
[1]. They are applied in a wide range of applications in
transportation planning and traffic engineering, in particular
in testing new vehicle systems for automated cars [2]. A
key issue related to automated driving is the communication
between automated vehicles (AVs) and vulnerable road users
(VRUs), because, as reviewed in [3], VRU accidents are often
associated with possible occlusion, unexpected trajectories or
difficulties of the VRU in threat perception. Communication
as an enhancement to human perception in road traffic offers
a possibility to improve safety in the VRU-to-AV interaction.
Several studies present human-machine interfaces (HMI) as
a communication method [4], [5], however, the majority of
HMI studies mainly focus on pedestrian-to-AV interaction.
The investigation with other user groups such as bicyclists
is also required for a valid HMI-concept evaluation [6]. For
complex traffic scenarios including multiple interacting traf-
fic participants, conventional single-seat driving simulators
(SSDS) reach their limits [7]. The behavior and interaction
of several human road users can be investigated with higher
validity when test subjects interact directly with each other in
a virtual environment. Individual simulators must therefore
*This work is part of the research project @CITY Automated Cars
and Intelligent Traffic in the City. The project is supported by the Federal
Ministry for Economic Affairs and Energy of Germany (BMWi), based on
a decision taken by the German Bundestag, grant number 19A17015B.
1Chair of Traffic Engineering and Control, Technical University of
Munich (TUM), 80333 Munich, Arcisstrae 21, Germany.
2CARISSMA, Technical University of Applied Sciences Ingolstadt, Es-
planade 10, 85049 Ingolstadt, Germany.
be linked to form multiple-seat driving simulators (MSDS).
In this paper, a concept of a coupled driving simulator
is presented to study the interaction between (automated)
vehicles and bicyclists. First, an overview of the state of the
art in coupled driving simulation is given (II), followed by
the explanation of the AV and bicycle simulator components
and their interactions (III). The simulator is tested in a proof-
of-concept study with 16 simulation runs. The results of a
survey, regarding the overall impressions of the simulation
and the interaction with the simulators, are presented in IV.
The main innovation of this setup is the investigation of
bicycle-to-AV interaction including mobile devices as HMI
components for both traffic participants. They serve on the
one hand as an information source, and on the other as
devices for influencing the driving behavior of the simulated
automated vehicle.
II. STATE OF THE ART OF COUPLED DRIVING
SIMULATIONS
Coupled simulations have been used for different appli-
cations in the past. One application field is driver training
[8]. Coupled simulators can provide a great advantage in
driver training compared to SSDS. The drivers being trained
drive not only in defined scenarios, but also have to react
to the driving behavior of other road users. Also, the driv-
ing instructor can deliberately create critical and dangerous
driving situations to better prepare the student for a real test
drive. Another application field, in which coupled simulation
is used, is engineering research and development [8]. In
this field, coupled simulators are applied to test new vehicle
systems, such as advanced driver-assistance systems (ADAS)
[9], [10]. Another example is the evaluation of HMI-concepts
that investigate the communication of automated vehicles
and pedestrians [1]. Besides the thematic categorization of
coupled driving simulator studies, they can be classified by
simulated traffic participants as well as hard- and software
components.
A. Simulated traffic participants in coupled driving simula-
tor studies
In numerous studies [10]–[18], among others, car-car inter-
action was investigated. The number of interacting vehicles
varies from a minimum two up to five or even more cars.
Since the year 2015, work can be found in which VRUs are
taken into account [1], [19]–[22]. The vehicle composition
always consists of an (automated) passenger car and a VRU
(motorcycle, bicycle or pedestrian), with pedestrian-to-car in-
teraction predominating. Bicycles are presented in two papers
with the focus on traffic safety [21], [23]. The bicycle-to-
AV interaction is underrepresented in transportation research
compared to pedestrian-to-AV interaction, although bicycle
traffic is an important transportation mode of urban traffic.
B. Hardware and Software of coupled driving simulators
Hardware and software components play a major role in
driving simulations. Little commercial software exists that
enable MSDS. In several studies only “SILAB” is mentioned
as commercial software [13], [18]–[20]. For the majority of
studies, custom software solutions are applied. These are
often based on the Unity game engine, either standalone or
in combination with other frameworks [1], [14]–[16], [23].
The hardware used for the simulators varies greatly across
studies from high-fidelity simulators to simple setups with
computer screens, steering wheel and pedals to simulate a
passenger car. With the increasing number of involved traffic
participants, the complexity of the individual simulators
decreases. The fixed base bicycle simulator in [23] detects
steering angle and speed while riding a real bicycle. Besides
the driving simulators themselves, additional hardware to
monitor the test subjects, for example eye tracking solutions,
are installed in some cases [7], [21]. A coupled driving
simulation contains many different hardware and software
solutions. They all must work together efficiently to produce
a high quality driving simulation. The combination of all
these components is a major challenge in implementing
coupled driving simulators [7]. In order to evaluate coupled
simulator systems and guarantee system resilience, four
requirements are defined in [8]: systematic approach, open
system interface, configurability principle and modularity
principle. These requirements represent an evaluation and
implementation basis for coupled driving simulator systems.
The literature review indicates that there is a lack of cou-
pled simulators that enable the investigation of HMI concepts
between AVs and bicyclists. Also little work exists regarding
HMI concepts for bicyclists, as described in [24]. Due to
the trends of vehicle automation [25] and the promotion
of sustainable modes of transport [26], [27], enabling safe
encounter between AVs and bicycles will be a crucial point
in future urban traffic.
III. METHODOLOGY
A. Simulator Study Design
In order to test the simulator setup, a proof-of-concept
study with 16 simulation runs and two participants (13
female, 19 male - Age group 18-24: 12, 25-39: 19, 40-
59: 1) each is conducted. During the study, an AV and its
passenger, and the bicyclist interact in a virtual environment.
The AV drives on a predefined path, while the bicyclist can
move freely in the virtual city model. The communication
application includes a navigation mode, which on the one
hand informs the AV-passenger about upcoming maneuvers
of the AV, and on the other hand guides the bicyclist through
the city network. At the intersection points of the vehicles’
Fig. 1: Conflict situation including bicyclist and AV
Fig. 2: Trigger points for the AV within an investigation
scenario: Pre-information (PE), Decision phase, virtual stop
line and the exit scenario trigger
routes, the road users have to interact (see Fig. 1). These
conflict points are chosen specifically to investigate certain
traffic scenarios. When approaching a conflict point, we de-
fine three interaction types, depending on the priority rules at
the conflict point and the information displayed on the HMI:
(1) the default type, where conventional traffic rules apply,
(2) the AV deciding autonomously about the traffic rules,
and (3) the AV-passenger deciding about the traffic rules. If
a non-default scenario is imminent, the application gives the
HMI user a pre-information about the upcoming scenario. If
the priority decision is up to the AV-passenger (option 3),
the test subject can decide about priority rules and is then
informed about the AV’s driving behavior during the conflict
scenario (see Fig. 2). If the AV decides autonomously about
priority rules in a traffic scenario (option 2), only information
about the vehicle’s behavior during the scenario is displayed.
The bicyclist also gets the pre-information followed by an
instruction on how to behave in the scenario (priority given
or taken). After leaving the traffic scenario, the application
switches back in navigation mode. The decisions of the AV or
its passenger are synchronised with the simulation. Depend-
Fig. 3: Coupled simulator setup (Left: AV simulator. Right: Bicycle simulator)
ing on the current decision, the AV behaves differently and,
for example, reduces the speed when the bicyclist is given
priority. After each simulation run, the study participants had
to fill out a survey. The results regarding the behaviour of
the vehicles in the virtual environment and the interaction
with the simulator are presented in IV.
B. Hardware
The coupled driving simulator consists of two parts: the
AV simulator and the bicycle simulator (see Fig. 3). The
AV simulator is composed of a PC with 3 monitors, a
speaker and a tablet as a communication device. Optional
are a steering wheel and pedals if the study includes a
non-automated car. The bicycle simulator requires a more
elaborate setup. For this purpose, the bicycle simulator of
the Chair of Traffic Engineering and Control at the Technical
University of Munich was used [28], [29]. This simulator is
implemented with a real bicycle that is located in front of a
television including a sound system. The steering movements
of the bicycle are measured by a magnetic rotary encoder,
mounted on a rotating plate on which the front wheel is
placed. The speed of the bicycle is calculated by an infrared
sensor which counts the rotations of a metal cylinder driven
by the bicycle’s rear wheel. Since only the rotation of the
rear wheel is measured, only the rear brake is functional
in this simulator setup. The sensor data is processed via an
Arduino Uno microcontroller which sends the data to a com-
munication port of a PC. The hand signals of the bicyclist
are detected via a depth camera. Moreover, a smartphone as
communication device is mounted on the bicycle handlebar.
The two driving simulators are located in different rooms
and connected via internet cable. For the communication
application, PC-1 acts as server for both client devices, the
tablet and the smartphone. During the entire communication,
the PCs and the mobile devices are connected to a virtual
private network (VPN). It is also possible to link the PCs via
a direct LAN connection. Both test subjects are monitored
with cameras during the simulation in order to intervene
when problems occur.
C. Software
To realize the coupled real-time simulator study, efficient
software solutions are required. The core components of the
simulation are the virtual environment, the visualisation and
control of the road users, the networking solution and the
communication via the web application. The information
flow between the components is displayed in a layer structure
in Fig. 4. Based on the study design above, the single layers
are explained in the following.
Unity game engine
The Unity game engine is the main part of the software
concept [30]. All information streams flow together or
distribute here. The virtual environment is built up and
the vehicle models can be controlled within the game
Fig. 4: Software layer structure of the coupled simulator
engine. This includes for the bicyclist the movement of
the vehicle as well as the character animation with hand
signals. The simulator input parameters (steering angle
and speed) are processed to control the virtual bicycle
model. The movement of the AV is controlled by an
algorithm that adjusts the speed of the car in such a way
that both road users enter the conflict situations at the
same time. When the AV is in an investigation scenario,
the communication application controls the speed input.
Thus, a two-way communication between game engine
and web application is realized.
A benefit of game engines used for coupled driving
simulation is the know-how of the gaming industry
in multiplayer games [1]. For this coupled driving
simulator, the built-in networking solution is applied.
It handles low level networking tasks and enables the
traffic researcher to work on the networking at a higher
abstraction level.
Web application
The web application provides the mobile device user in-
formation about navigation, traffic situations or vehicle
behavior. It also enables the AV-passenger to interact
with the vehicle and decide at certain situations about
priority rules at a conflict point. In between investigation
scenarios, the application is in navigation mode. It
provides the AV-passenger information about the AV’s
behavior or guides the bicyclist through the virtual city
network. Upon entering an investigation scenario, the
application conveys basic information about the traffic
situation and the upcoming conflict with another vehi-
cle. Also, real-time information about the distance to the
conflict point is displayed. After the priority decision to
resolve the conflict situation, corresponding information
is provided to the users. The web application also
communicates this decision to the Unity game engine
component, in order to adjust the AV’s driving behavior.
The communication between the web application and
other software components is realized using the internet
protocol HTTP. Access to the application is provided via
a representational state transfer (REST) programming
interface [31]. Thus, the application only has certain
states, which can be triggered from an external software.
During the development phase, it was a benefit to
run the application with web technology, because it is
platform- and device-independent. A version for native
Android support is under development.
HMI-Layer
The human-machine interface layer is represented by
the from the web application provided information to
the mobile device user. Not only visual information in
text or image form is displayed, but also audio signals
are used to encode the message information by type.
For example, warning messages have a unique signal
tone.
For simulation purpose, a mobile phone for the bicyclist
and a tablet for the AV-passenger are used as hardware
components. In reality, this concept is applicable for
the bicyclist, because it is already common to use a
mobile phone as a navigation device. Such a navigation
application can then be extended with conflict infor-
mation functionalities. In an automated vehicle it is
conceivable that the automation HMI (aHMI) in the
future is represented by a tablet. Nowadays it is practice
that aHMIs are integrated in the instrument cluster in
the windshield, a monitor on the center console or head-
up displays [32]. A more detailed description of the
communication application can be found in [24].
User-Layer
The test subject must be able to process the received
information correct in content in adequate time. It
has to evaluate the input from different information
channels: the simulation world and the HMI device.
Based on the bicyclist subject’s perception, it adjusts the
driving behavior and decisions. The AV-passenger is not
required to perform driving actions. Rather, the focus is
on user acceptance and trust in the automated system,
and for this study design, on the priority decisions of
the AV-passenger.
Input-Layer
The input layer represents the input from the physical
driving simulators. The bicycle simulator provides the
simulation with information about the steering angle and
rear wheel’s rotation. The raw values are adjusted to
calibrate the sensitivity of the system input. The hand
signs are detected by a depth camera with an underlying
machine learning algorithm [33]. It detects whether a
hand signal is being given based on skeleton points of
Fig. 5: Bicyclist evaluation of the simulation impressions and
interaction with the simulator
the bicyclist. This information is synchronized to the
animations in the game engine.
The input from the AV-passenger is handled by the web
application. The application transfers the decision about
priority rules to the game engine.
During the simulation data is collected from the study
participants and saved in two CSV files. The game engine
writes one file with information about position, rotation and
the use of blinkers and hand signs. The web application saves
time and type of when a HMI screen is triggered. Moreover,
the test subjects had to fill in one short survey about the
performance of the web application and the safety perception
right after each investigation scenario. They also answered an
extensive survey including demographic questions, questions
on their driving experience, the overall impressions of the
simulator study and the interaction with the simulators after
the whole simulation run. The results of the survey part
regarding the overall impressions and interaction with the
simulators are presented below.
IV. RESULTS
For the evaluation of the simulator setup, we analyzed
questions from the survey the test subjects filled in after
a simulation run (see Figs. 5 and 6). The content of these
questions concerns the impressions of the simulation and the
interaction with the simulators. The questions regarding the
individual simulators differ from bicyclist to AV-passenger.
Each parameter could be evaluated from 1 (very unrealistic)
to 5 (very realistic).
The overall impression of both test subjects tends to
be positive (Bicycle: 3.20, AV: 3.75). The same trend is
found for the parameter 3D graphics (Bicycle: 3.13, AV:
3.81). Some other evaluations differ significantly between the
bicyclist and the AV-passenger. The behavior of other traffic
Fig. 6: AV-passenger evaluation of the simulation impres-
sions and interaction with the simulator
participants was rated with 3.93 points from the bicyclists.
On average 1.00 points higher than the AV-passenger’s rating.
Because only these two traffic participants, the bicycle and
the AV, were in the simulated environment, the bicycles
behavior is rated more poorly than the AV’s behavior. The
reason for that may be that the tilt angle in a curve wasn’t
visualized in the simulation to avoid motion sickness for
the bicyclist subject. A possible solution to this would be
different visualization for each test subject. The bicyclist
will see a model synchronized to the simulator to ensure
intuitive control; the AV a highly detailed, realistically-
behaving model. Besides the tilt angle parameter, the head
movement and therefore eye contact as a common commu-
nication pattern between cars and bicyclists [34] was not
represented in the simulation at all. The sound was rated with
3.80 points for the bicyclist and with 2.31 for the AV. Some
AV-passenger subjects noted that they did not recognize any
sound. It seems that they got used to the sound scenery in the
simulated world, which did not include many sound sources.
The field of view evaluation is highly dependent on the
simulator setup. It was rated at 2.67 points by the bicyclists
and 4.25 point by the AV-passengers. For the bicyclist only
one screen was available. The natural field of view of about
50° is artificially extended by modifying the recording angle
of the simulated camera to 90°. Much less, compared to the
165° field of view for the AV-passenger. In future studies,
more screens or virtual reality technology will be applied
to increase the bicyclist’s field of view. The remaining pa-
rameters are vehicle dependent. The bicyclists evaluated the
speed, braking and pedaling behavior somewhat favorably.
The steering behavior was evaluated as less realistic, with
2.67 points. Two influencing factors are the already discussed
tilt angle in a curve and the restricted field of view. The
AV-passengers rated parameters regarding the AV’s driving
behavior (speed, acceleration, steering) tending unrealistic.
This is mainly caused by the algorithm that controls the
AV’s acceleration and deceleration, as discussed with study
participants. The AV in the simulation must adjust to the
bicyclists speed to arrive at the same time with the bicycle
at the investigation scenarios. This can lead to unnatural
speed choice and acceleration behavior. This behavior is only
recognized in between investigation scenarios, and is thus
not critical for the scenarios themselves. The 3D model of
the car was rated very realistic. For the model, a slightly
modified version of the car model used in [1] was applied.
This detailed model includes for example mirrors, lights
and a speedometer. For the investigation of traffic scenarios
with human-human interaction, the parameter “behavior of
other traffic participants” is especially important. The bicycle
model will be improved with regard to important communi-
cation patterns like hand signals and eye contact [34]. The
networking solution can here solve the problem of road-user-
dependent visualization.
V. CONCLUSIONS
A coupled driving simulator including an automated or
non-automated vehicle and a bicycle is implemented. It en-
ables traffic participants to interact in a virtual environment.
Mobile devices are included to represent HMI components
in the simulation. The visualization and interaction with the
HMI is implemented with a web application. It is possible to
provide the HMI user with information linked to the driving
simulation and also let the user decide about priority rules in
specific conflict scenarios. The mobile devices can thus also
be used as input devices to influence the vehicles behavior.
Moreover, the following points will be implemented in the
future:
Interface to SUMO
In the presented study, only two vehicles interact in
the simulation. For more realistic and complex traffic
scenarios, an interface to the traffic simulation software
SUMO will be implemented [35].
Full body motion tracking of the bicyclist
The movement of all body parts, especially the arms for
realistic hand signals, will be animated and compared
to the discrete animations currently used (give a hand
signal or not). This increases the simulation validity of
hand signals and enables communication methods like
eye contact.
ACKNOWLEDGMENT
This work is part of the research project @CITY
Automated Cars and Intelligent Traffic in the City. The
project is supported by the German Federal Ministry for
Economic Affairs and Energy (BMWi), based on a decision
taken by the German Bundestag, grant number 19A17015B.
The authors are solely responsible for the content of this
publication.
REFERENCES
[1] P. Bazilinskyy, L. Kooijman, D. Dodou, and J. de Winter, “Coupled
simulator for research on the interaction between pedestrians and
(automated) vehicles,” 09 2020.
[2] J. Barcel´
o, Fundamentals of traffic simulation. Springer, 2010, vol.
145.
[3] F. Denk, W. Huber, P. Brunner, and R. Kates, “The role of perceptual
failure and degrading processes in urban traffic accidents: a stochastic
computational model for virtual experiments,” 09 2020, pp. 1–7.
[4] A. Rouchitsas and H. Alm, “External humanmachine interfaces for au-
tonomous vehicle-to-pedestrian communication: A review of empirical
work,” Frontiers in Psychology, vol. 10, 2019.
[5] P. Bazilinskyy, D. Dodou, and J. de Winter, “Survey on ehmi concepts:
The effect of text, color, and perspective,” Transportation Research
Part F: Traffic Psychology and Behaviour, vol. 67, pp. 175–194, 2019.
[6] D. Dey, A. Habibovic, A. Lcken, P. Wintersberger, B. Pfleging,
A. Riener, M. Martens, and J. Terken, “Taming the ehmi jungle:
A classification taxonomy to guide, compare, and assess the design
principles of automated vehicles’ external human-machine interfaces,
Transportation Research Interdisciplinary Perspectives, vol. 7, p.
100174, 2020.
[7] B. D. Sawyer and P. Hancock, “Development of a linked simulation
network to evaluate intelligent transportation system vehicle to vehicle
solutions,” Proceedings of the Human Factors and Ergonomics Society
Annual Meeting, vol. 56, no. 1, pp. 2316–2320, 2012.
[8] K. Abdelgawad, J. Gausemeier, R. Dumitrescu, M. Grafe, J. Stcklein,
and J. Berssenbrgge, “Networked driving simulation: Applications,
state of the art, and design considerations,” Designs, vol. 1, no. 1,
2017.
[9] T. Bando and T. Shibata, “Networked driving simulator based on
sigverse and lane-change analysis according to frequency of driving,
in 2012 15th International IEEE Conference on Intelligent Transporta-
tion Systems, 2012, pp. 1608–1613.
[10] A.-U.-H. Yasar, Y. Berbers, D. Preuveneers, and A. Jameel, A
computational analysis of driving variations on distributet multiuser
driving simulators, 01 2008.
[11] P. Hancock and S. de Ridder, “Behavioural accident avoidance science:
understanding response in collision incipient conditions,” Ergonomics,
vol. 46, no. 12, pp. 1111–1135, 2003.
[12] M. Houtenbos, “Expecting the unexpected,” Ph.D. dissertation, Delft
University of Technology, 2008.
[13] C. Maag, D. Muhlbacher, C. Mark, and H.-P. Kruger, “Studying effects
of advanced driver assistance systems (adas) on individual and group
level using multi-driver simulation, IEEE Intelligent Transportation
Systems Magazine, vol. 4, no. 3, pp. 45–54, 2012.
[14] K. Gajananan, A. Nantes, M. Miska, A. Nakasone, and H. Prendinger,
“An experimental space for conducting controlled driving behavior
studies based on a multiuser networked 3d virtual environment and the
scenario markup language,” IEEE Transactions on Human-Machine
Systems, vol. 43, no. 4, pp. 345–358, 2013.
[15] H. Prendinger, K. Gajananan, A. Bayoumy Zaki, A. Fares, R. Mole-
naar, D. Urbano, H. van Lint, and W. Gomaa, “Tokyo virtual living
lab: Designing smart cities based on the 3d internet,” IEEE Internet
Computing, vol. 17, no. 6, pp. 30–38, 2013.
[16] H. Prendinger, M. Miska, K. Gajananan, and A. Nantes, A cyber-
physical system simulator for risk-free transport studies,” Computer-
Aided Civil and Infrastructure Engineering, vol. 29, no. 7, pp. 480–
495, 2014.
[17] K. Oeltze and C. Schiel, “Benefits and challenges of multi-driver
simulator studies,” IET Intelligent Transport Systems, vol. 9, no. 6,
pp. 618–625, 2015.
[18] D. Muehlbacher, The Multi-Driver Simulation: A Tool to Investigate
Social Interactions Between Several Drivers. Wiesbaden: Springer
Fachmedien Wiesbaden, 2018, pp. 379–391.
[19] C. Lehsing, A. Kracke, and K. Bengler, “Urban perception - a cross-
correlation approach to quantify the social interaction in a multiple
simulator setting,” in 2015 IEEE 18th International Conference on
Intelligent Transportation Systems, 2015, pp. 1014–1021.
[20] S. Will, A New Approach to Investigate Powered Two Wheelers’
Interactions with Passenger Car Drivers: the Motorcycle Car Multi-
Driver Simulation. Wiesbaden: Springer Fachmedien Wiesbaden,
2018, pp. 393–402.
[21] J. Warner, D. S. Hurwitz, C. M. Monsere, and K. Fleskes, A
simulator-based analysis of engineering treatments for right-hook
bicycle crashes at signalized intersections,” Accident Analysis &
Prevention, vol. 104, pp. 46–57, 2017.
[22] Y. E. Song, C. Lehsing, T. Fuest, and K. Bengler, “External hmis
and their effect on the interaction between pedestrians and automated
vehicles,” in Intelligent Human Systems Integration, W. Karwowski
and T. Ahram, Eds. Cham: Springer International Publishing, 2018,
pp. 13–18.
[23] D. Noyce, J. Kearney, S. Chakraborty, Y. Jiang, and K. Santiago-
Chaparro, “Multi-modal distributed simulation combining cars, bicy-
clists, and pedestrians,” 09 2018.
[24] J. Lindner, G. Grigoropoulos, A. Keler, P. Malcolm, F. Denk, P. Brun-
ner, and K. Bogenberger, “A mobile application for resolving bicyclist
and automated vehicle interactions at intersections,” 06 2022.
[25] A. Mahdavian, A. Shojaei, S. Mccormick, T. Papandreou, N. Eluru,
and A. A. Oloufa, “Drivers and barriers to implementation of con-
nected, automated, shared, and electric vehicles: An agenda for future
research,” IEEE Access, vol. 9, pp. 22195–22 213, 2021.
[26] J. Pucher, C. Komanoff, and P. Schimek, “Bicycling renaissance in
north america?: Recent trends and alternative policies to promote
bicycling,” Transportation Research Part A: Policy and Practice,
vol. 33, no. 7, pp. 625–654, 1999.
[27] C. European, “The european green deal,” https://eur-
lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:52019DC0640,
2019.
[28] A. Keler, J. Kaths, F. Chucholowski, M. Chucholowski, G. Grig-
oropoulos, M. Spangler, H. Kaths, and F. Busch, A bicycle simu-
lator for experiencing microscopic traffic flow simulation in urban
environments, in 2018 21st International Conference on Intelligent
Transportation Systems (ITSC), 2018, pp. 3020–3023.
[29] H. Kaths, A. Keler, J. Kaths, and F. Busch, Analyzing the behavior
of bicyclists using a bicycle simulator with a coupled sumo and dyna4
simulated environment, 08 2019.
[30] J. K. Haas, “A history of the unity game engine, 2014.
[31] R. T. Fielding, Architectural styles and the design of network-based
software architectures”; doctoral dissertation,” 2000.
[32] K. Bengler, M. Rettenmaier, N. Fritz, and A. Feierle, “From hmi to
hmis: Towards an hmi framework for automated driving, Information,
vol. 11, no. 2, 2020.
[33] P. Malcolm, G. Grigoropoulos, A. Keler, H. Kaths, and K. Bogen-
berger, Analysis of bicyclist communication in a simulator environ-
ment,” 01 2021.
[34] B. F ¨
arber, Communication and Communication Problems Between
Autonomous Vehicles and Human Drivers”, bookTitle=”Autonomous
Driving: Technical, Legal and Social Aspects. Berlin, Heidelberg:
Springer Berlin Heidelberg, 2016, pp. 125–144.
[35] P. A. Lopez, M. Behrisch, L. Bieker-Walz, J. Erdmann, Y.-P. Fl¨
otter¨
od,
R. Hilbrich, L. L¨
ucken, J. Rummel, P. Wagner, and E. Wießner,
“Microscopic traffic simulation using sumo,” in The 21st IEEE In-
ternational Conference on Intelligent Transportation Systems. IEEE,
2018.
ResearchGate has not been able to resolve any citations for this publication.
Conference Paper
Full-text available
In order to facilitate safe interactions between automated vehicles (AVs) and vulnerable road users (VRUs) such as bicyclists, we present a communication application for mobile devices that allows an AV or its passenger and a bicyclist to interact in certain traffic scenarios. At the intersection, the AV or its passenger can change the existing right-of-way rules to prioritise the ego-vehicle or the bicyclist. In a coupled driving simulator in which these two road users can interact, 16 proof-of-concept experiments are conducted. It is found that the perceived safety at conflict points can be increased through the use of the application. An investigation of the user data provides insights into the AV passengers' decision types and duration in the scenarios studied. Moreover, the simulation results are used to revise and further develop the application concept.
Article
Full-text available
Several converging trends appear to reshape the way citizens and goods move about. These trends are social, including urbanization and population growth, and technological, such as increased automation and connectivity. All these factors influence the market for connected, automated, shared and electric (CASE) vehicles, which presents many opportunities and challenges. The pace of the shift to a profoundly penetrated market for CASE vehicles is far from secure. Such transformation depends on the development of technologies, consumer attitudes, and policies. An expanding body of research has investigated the potential social and behavioral results of deploying CASE vehicles. However, most academic literature to date concentrates on technological issues linked to these vehicles. There are several teams from federal and state agencies, OEMs, academia, startups, and consortiums working on this complex subject. This study investigates several academic papers, as well as federal and industry reports, considering all the stakeholders mentioned above. Its aim is to present a comprehensive picture of the implementation barriers and drivers of CASE vehicle usage and provide suggestions to solve them. The findings confirm that several issues are currently affecting the implementation of CASE vehicles on the road. Although there have been significant partnerships and collaborations between CASE vehicle stakeholders, namely technology companies, federal-state agencies, and academic scholars, considerable work is still required to solve the remaining barriers facing CASE-related technologies. This would enable decision-makers to create effective policies for future transportation networks and increase the speed of CASE vehicle market penetration to enhance road network’s level of service.
Poster
Full-text available
Urban roads are a dynamic environment in which formal and informal communication are crucial. In order for autonomous vehicles to operate effectively in such an environment, they must be able to deal with this complexity. In particular, an understanding of the communication patterns of bicyclists is crucial because of their vulnerability and lack of standardized indicators, which allows greater variability and subtlety in the way they communicate. In this study, participants rode a bicycle simulator through various urban traffic scenarios in which explicit and implicit communication behaviors are expected. Participants were recorded with a depth camera, and a markerless motion capture technique was used to record their movements in three dimensions. From this data, hand signal, head movement, and leaning events were extracted, analyzed, and compared. Analyses showed that even in clearly regulated scenarios, not all participants performed a hand signal, with between 80 and 95 percent of participants signaling before turning at a four-way intersection and approximately 60 percent before changing lanes. Participants were also significantly more likely to glance over their left shoulder preceding a left lane-change or left turn compared to other scenarios. The arm shape while performing a hand signal was found to be almost entirely governed by individual preference, rather than scenario. Based on these results, it is clear that cyclist communication behavior frequently does not adhere strictly to traffic regulations and varies from person to person. However, implicit cues such as head movements can be used to supplement behavior prediction models in certain situations.
Article
Full-text available
There is a growing body of research in the field of interaction between automated vehicles and other road users in their vicinity. To facilitate such interactions, researchers and designers have explored designs, and this line of work has yielded several concepts of external Human-Machine Interfaces (eHMI) for vehicles. Literature and media review reveals that the description of interfaces is often lacking in fidelity or details of their functionalities in specific situations, which makes it challenging to understand the originating concepts. There is also a lack of a universal understanding of the various dimensions of a communication interface, which has impeded a consistent and coherent addressal of the different aspects of the functionalities of such interface concepts. In this paper, we present a unified taxonomy that allows a systematic comparison of the eHMI across 18 dimensions, covering their physical characteristics and communication aspects from the perspective of human factors and human-machine interaction. We analyzed and coded 70 eHMI concepts according to this taxonomy to portray the state of the art and highlight the relative maturity of different contributions. The results point to a number of unexplored research areas that could inspire future work. Additionally, we believe that our proposed taxonomy can serve as a checklist for user interface designers and researchers when developing their interfaces.
Article
Full-text available
During automated driving, there is a need for interaction between the automated vehicle (AV) and the passengers inside the vehicle and between the AV and the surrounding road users outside of the car. For this purpose, different types of human machine interfaces (HMIs) are implemented. This paper introduces an HMI framework and describes the different HMI types and the factors influencing their selection and content. The relationship between these HMI types and their influencing factors is also presented in the framework. Moreover, the interrelations of the HMI types are analyzed. Furthermore, we describe how the framework can be used in academia and industry to coordinate research and development activities. With the help of the HMI framework, we identify research gaps in the field of HMI for automated driving to be explored in the future.
Technical Report
Full-text available
A collaborative project between the University of Wisconsin-Madison and the University of Iowa was conducted to understand the feasibility of conducting driving and pedestrian simulator experiments that involve multiple agents (humans) in the simulation. As part of the collaborative project, a distributed driving simulation environment was created that can be used to conduct driving simulation experiments with multiple human agents participating in the simulation as a pedestrian or as a driver. The distributed simulation environment was created using the Unity game engine. A pedestrian can participate in the experiment using a virtual reality headset, while the driver can participate using a regular computer connected to a screen and a steering wheel control system. The distributed nature of the environment, ie, the ability to have subjects be part of the same experiment from geographically distant locations, is achieved by relying on the built-in networking functionality provided by the Unity game engine. As will be described in the report, the simulation environment was created by relying on the tools provided by the integrated development environment used by the Unity game engine and by creating supplemental modules created to support common tasks in driving simulation experiments.
Conference Paper
Full-text available
Driving simulators are regarded as valuable tools for human factors research on automated driving and traffic safety. However, simulators that enable the study of human-human interactions are rare. In this study, we present an open-source coupled simulator developed in Unity. The simulator supports input from head-mounted displays, motion suits, and game controllers. It facilitates research on interactions between pedestrians and humans inside manual and automated vehicles. We present results of a demo experiment on the interaction between a passenger in an automated car equipped with an external human-machine interface, a driver of a manual car, and a pedestrian. We conclude that the newly developed open-source coupled simulator is a promising tool for future human factors research.
Article
Full-text available
Interaction between drivers and pedestrians is often facilitated by informal communicative cues, like hand gestures, facial expressions, and eye contact. In the near future, however, when semi- and fully autonomous vehicles are introduced into the traffic system, drivers will gradually assume the role of mere passengers, who are casually engaged in non-driving-related activities and, therefore, unavailable to participate in traffic interaction. In this novel traffic environment, advanced communication interfaces will need to be developed that inform pedestrians of the current state and future behavior of an autonomous vehicle, in order to maximize safety and efficiency for all road users. The aim of the present review is to provide a comprehensive account of empirical work in the field of external human–machine interfaces for autonomous vehicle-to-pedestrian communication. In the great majority of covered studies, participants clearly benefited from the presence of a communication interface when interacting with an autonomous vehicle. Nevertheless, standardized interface evaluation procedures and optimal interface specifications are still lacking.
Conference Paper
Full-text available
Operational behavior models are used in traffic simulations to represent the subconscious, short-term decisions made by road users to respond to other road users, the infrastructure and traffic control measures. Calibration and validation of these models can be achieved using observed trajectory data from real road users. For lane bound traffic, it is assumed that road users intend to follow a given lane with a certain desired speed across the intersection. Any deviation from this planned path is in response to other road users or the environment. It is difficult, however, to identify and separate the desired movement of more flexible road users that do not follow lane disciple, such as bicyclists, from movements made as a reaction to other road users or obstacles. This can lead to poor calibration of operational behavior models and unrealistic behavior in the simulation. Tactical behavior models recreate the conscious decisions made on a time horizon of seconds to minutes to cope with the immediate traffic situation. As such, tactical behavior models are responsible for selecting the planned path across an intersection. Here, SUMO is coupled with the simulation software DYNA4 to create a simulated road environment for a bicycle simulator. Trajectories observed in reality are displayed as potential prescribed pathways across the simulated intersection. Participants in the simulator study are instructed to select and follow one of the prescribed pathways as closely as possible while responding naturally to other road users and obstacles in the environment. The resulting trajectory data is used to calibrate existing operational and tactical path finding behavior models for bicyclists at signalized intersection.