Presentation

· Contributors · Organizations · Search Program · Flagged · Happening Now

Game-theoretic Strategies in Mixed Traffic: A Simulator Study on the Impacts of Automated and Human Driving Styles

SessionST7: Topics in Automated and Human Driving

DescriptionAs automated driving technology advances, human-driven vehicles (HVs) will continue sharing the road with automated vehicles (AVs). Traffic systems are shifting toward mixed human-machine cooperation. This transition introduces two cooperation patterns: vertical cooperation between drivers and vehicle automation (e.g., takeovers in conditionally AVs) and horizontal cooperation between HVs and AVs. However, discrepancies between AV behavior and human expectations may cause misunderstandings, leading to inconsistent driver responses and unpredictable driving patterns, increasing collision risks. Understanding HV-AV interactions is crucial for assessing driver adaptation to automation. Research on human decision-making helps identify key factors influencing safe and efficient interactions, informing AV design for more predictable cooperation. Although studies have examined HV-AV interaction through surveys, simulators, and naturalistic driving data, gaps remain in understanding driver adaptation in mixed-traffic environments. Our research contributes in three ways: First, analyzing diverse driving behaviors in HV-AV interactions,
addressing the assumption that AVs always follow conservative driving patterns; Second, employing a rule-based game-theoretic framework to model decision-making in parallel and head-on traffic scenarios, considering both longitudinal and lateral control; Third, applying trajectory-based clustering to link driving styles to trajectory variations, improving AV modeling and cooperative driving strategies.

Contributors

Yutong Zhang

University of Pittsburgh

Shiqi Wu

Chalmers University of Technology

University of Gothenburg

Danneil Mubbala

University of Pittsburgh