Lane-Change Behavior of Human Drivers in Mixed Traffic: A Driving Simulator Study

Connected and autonomous vehicles (CAVs) and human-driven vehicles (HDVs) will inevitably coexist on roads in the near future, creating a mixed-traffic environment. This coexistence introduces safety-critical interactions during human drivers’ lane changes, posing key challenges for CAV operations and potentially undermining traffic safety. Hence, it is necessary to understand human drivers’ lane-change behavior in mixed traffic. This study analyzes the lane-change behavior of human drivers in HDV-only and mixed-traffic contexts. It considers proactive CAV control strategies that execute maneuvers before HDV lane changes and passive CAV control strategies to manage the induced oscillations after HDV lane changes. Multiple surrogate safety metrics are adopted to analyze lane-change performance from three aspects: reaction time to potential collisions, risk of imminent collisions, and braking response. A novel safety performance framework is proposed to comprehensively integrate these safety metrics into a unified measure, partly accounting for their interdependencies. Human drivers’ lane-change behavior is analyzed using trajectory data from driving simulator experiments involving 72 participants. The study results illustrate that human drivers exhibit diminished safety performance (e.g., smaller minimum time headways and minimum time-to-collisions) when performing lane changes in front of CAVs in a CAV platoon, primarily attributed to the small vehicle spacings. Further, they exhibit significant novelty effects at their first contact with CAVs, resulting in impaired lane-change performance. However, they also illustrate a capability to learn from interactions with CAVs controlled by proactive control strategies and modify their lane-change behavior, which can enhance traffic safety. The study findings provide valuable insights on the evolution of human drivers’ lane-change behavior and the dynamics of CAV-HDV interactions in mixed traffic, illustrating practical implications for the development of CAV operation systems tailored to mixed traffic contexts.