Rosa P: A digital library for transportation research

U.S. Department of Transportation, Climate Change Center Climate Strategies That Work: Heavy-Duty Diesel Engine Retrofit and Replacement

NDE: FHWA Advanced Sensing Technology & Nondestructive Evaluation Program

Evaluating Candidate Projects for Seal Coat Stability

Consultant Support for Intelligent Compaction and Paver-Mounted Thermal Profiling Projects in 2024-2025

Implementation of Balanced Mixture Design in Missouri Test Sections with Modifiers [Summary]

Insights for the Future of Car Rental and Ridesharing: Driving Behavior Across Different Levels of Automation

• 
SJSU SAN JOSE STATE UNIVERSITY 
IOUGU 
210· 330• 
24D0 300° 
270" 
100 
60 
0 
Sp00d (mph) 
. · ......... ~ 
. . 
:;~~~i 
1MINETA TRANSPORTATION INSTITUTE 
Insights for the Future of Car Rental and Ridesharing: Driving 
Behavior Across Diferent Levels of Automation 
Pranav Meda Wei-Hsiang Lo Yue Luo, PhD Project 2427 
Aubrey Victoria Contreras Gaojian Huang, PhD February 2025 
Introduction 
Autonomous vehicles, such as those from Waymo, 
are reshaping car rental and ridesharing industries, 
suggesting a potential shift towards a unifed model 
of on-demand transportation. Tis shift in mobility 
brings new challenges, and understanding human 
behavior in response to diferent levels of vehicle 
automation and assistance features is crucial to safety. 
Tis study explores behavior, specifcally in driving 
performance, driving posture, and eye movement, 
among drivers when they drive in vehicles at diferent 
automation levels (Levels 0, 3, and 5). Te study 
also looks at how these behaviors vary with diferent 
assistance feature styles (risky and conservative modes) 
and in diferent driving activities (lane keeping and 
lane changing tasks). 
Study Methods 
Te study involved twelve participants (aged 21 to 
29, 4 males and 8 females) recruited to complete 
six driving trials via a driving simulator (MiniSim), 
each lasting about 6 minutes. Tese trials simulated 
driving scenarios that required lane keeping or lane 
changing in response to unexpected events, under 
one of three levels of automation—Level 0 (manual), 
Level 3 (conditional automation), and Level 5 (full 
automation)—and one of two assistance features— 
conservative (early indication/action) or risky (late 
indication/action). During the experiment, driving 
performance data, including speed (in mph) and 
steering wheel angles (in degrees), were recorded 
using the driving simulator. Driving posture data, 
such as fexion angles of the right wrist and right 
ankle, as well as body positioning of various body 
parts were captured using a Movella/Xsens motion 
capture system. Eye movement data, specifcally gaze 
coordination and pupil diameters, were tracked using 
a Pupil Core eye tracking system. Participants also 
completed pre-study, post study, and NASA-TLX 
questionnaires to survey their driving experience and 
attitudes toward automated vehicles. Te collected 
data were analyzed and presented to identify and 
compare changes in driving behaviors under diferent 
automation levels and assistance features. 
M I  N E  T  A  T  R  A N S P O R  T  A  T  I O N  I  N S T  I  T  U  T  E  transweb.sjsu.edu 
Project 2427   February 2025 M I  N E  T  A  T  R  A N S P O R  T  A  T  I O N  I  N S T  I  T  U  T  E  
 
      
 
 
.7..,.. 
. ,·::~'!~-~ 
~ ~ •. ·:,•,-~~~~ .... ·... . 
• 
I 
• • .......... 
-~!~~~=-~\ 
MINETA 
TRANSPORTATION 
INSTITUTE 
Findings 
Transitions between manual driving (Level 0), 
conditional automation (Level 3), and full driving 
automation (Level 5) revealed short-term behavior 
changes as drivers interacted with vehicles at diferent 
automation levels. In Level 0, where drivers had 
full control, stable control of speed and steering 
were maintained during lane keeping and lane 
changing. As vehicles transitioned to Level 3, where 
the vehicle completed driving tasks under normal 
conditions with drivers’ readiness still required, there 
was a noticeable increase in variability in speed and 
steering, especially under risky conditions where 
changes were more pronounced. At Level 5, with the 
vehicle fully in control, consistent speed and steering 
control were observed. Te study also noted changes 
in body posture that refected the driver’s adaptation 
to automation levels. Eye movement also shifted with 
automation levels. 
When drivers need to take control of a Level 
3 automated vehicle, they face challenges 
such as adjusting to changes in vehicle 
behavior and getting into the right posture to 
Policy Recommendations 
Considering the observed changes in driving 
performance, body posture, and eye movement across 
automation levels, there are three recommendations 
for vehicle design and driver training that can 
efectively address short-term behavioral changes. 
First, adaptive interface and alarm systems should be 
implemented to capture the driver’s attention across 
all automation levels. Tese systems should utilize 
multimodal cues—visual, auditory, and haptic—to 
ensure drivers can quickly and easily understand when 
intervention is needed, improving response time and 
accuracy. Secondly, the ergonomic design of vehicle 
controls should be enhanced to ease the physical and 
posture demands during transitions between diferent 
automation levels, which includes developing 
adjustable seating and steering mechanisms that adapt 
to the driver’s physical state. Additionally, training 
programs focusing on cognitive load management 
should be developed to facilitate drivers’ multitasking 
capabilities, complemented by workshops that 
increase driver awareness about the behavioral 
impacts when transitioning between diferent levels 
of automation. As autonomous vehicles become 
increasingly common, these considerations can help 
improve safety for everyone. 
About the Authors 
Pranav Meda is a master’s student in Human Factor/ 
Ergonomics at San José State University (SJSU), 
having earned a BS in Psychology from the University 
of California, Los Angeles. 
Aubrey Victoria Contreras is a master’s student in 
Human Factor/Ergonomics at SJSU, having earned a 
BA in Psychology from the University of California, 
Davis. 
Wei-Hsiang Lo is a master’s student in Human 
Factor/Ergonomics at SJSU, having earned a BA in 
Industrial design from Shih Chien University. 
Dr. Gaojian Huang is an Assistant Professor in the 
Industrial & Systems Engineering Department at 
SJSU with a PhD from Purdue University. 
Dr. Yue Luo is an Assistant Professor in the Industrial 
& Systems Engineering Department at SJSU with a 
PhD from the University of Florida. 
To Learn More 
For more details about the study, download the full 
report at transweb.sjsu.edu/research/2427 
MTI is a University Transportation Center sponsored by the 
U.S. Department of Transportation’s Ofce of the Assistant 
Secretary for Research and Technology and by Caltrans. Te 
Institute is located within San José State University’s Lucas 
Graduate School of Business. 
transweb.sjsu.edu M I  N E  T  A  T  R  A N S P O R  T  A  T  I O N  I  N S T  I  T  U  T  E  


Insights for the Future of Car Rental and Ridesharing: Driving Behavior Across Different Levels of Automation [Brief]

Michigan Department of Transportation Fleet Electrification Strategies

Ensuring Fair and Equitable Funding of Rural Transit in Georgia after the 2020 Census

Chemical Incident Reports: Events Reported to the CSB Under the Accidental Release Reporting Rule [Volume 1]

Not a member yet