California Intersection Decision Support: A Driver-Centered Approach to Left-Turn Collision Avoidance System Design

Currently, Federal and State governments are partnering with private industries and academia institutions to pursue the deployment of intersection decision support (IDS) and cooperative intersection collision avoidance systems (CICAS), which seek to combine infrastructure-based and vehicle-based functions to provide optimal solutions for roadway users. The overall (IDS) research plan was constructed to explore the requirements, tradeoffs assessment, and technology investigations necessary to define a Cooperative Intersection Collision Avoidance System. This report is the third report on the California PATH IDS research, and it focuses on two human factors studies which used the PATH instrumented Ford Taurus research vehicle to study driver behavior while making left turns. The goal of the first study was to observe drivers? intersection approaches and left-turn maneuvers in a mostly naturalistic setting. Twenty-three drivers (both male and female and younger and older) were recruited to drive the instrumented vehicle 10 times around an 8-block course in the city of Berkeley, CA, making a total of 40 left turns per driver. The instrumented vehicle recorded driver actions, such as approach speed, brake activation, steering inputs, and limited estimates of oncoming vehicle gap (and lag) acceptance. The goal of the second study at the California PATH Richmond Field Station (RFS) instrumented intersection was twofold. First, left-turn gap (or lag) acceptance was examined in an environment where gaps could be more accurately measured and tightly controlled. Second, the study introduced drivers to the concept of a left-turn Driver Infrastructure Interface (DII), a dynamic, no-left-turn sign, warning sign. Twenty drivers (both male and female and younger and older) were recruited to drive the instrumented vehicle through the RFS intersection 54 times with a single oncoming vehicle approaching. The vehicle approaches were timed to test the effects of different DII settings such as warning threshold onset timing on gap (lag) acceptance.