Vision-Based Identification and Application of the Leader Vehicle in Urban Environment

Comprehensive situational awareness is crucial to the effectiveness of advanced driver assistance systems (ADAS) and autonomous vehicles. This paper addresses a vision-based approach to identify the leader vehicle in complex urban environment, whose behavior is subsequently employed for the control purposes of the host vehicle. More specifically, the surrounding vehicles are detected and assigned to the corresponding driving lanes. Among those in the host lane, the one that most agrees with the host vehicle's driving task will be determined as the leader vehicle. Subsequently, the real-time, on-site and validated information of the leader vehicle, such as trajectory and velocity, is used to make the "stop" or "go" decisions as well as plan a safe and human-like local path and velocity profile. Experimental results in various typical but challenging urban traffic scenes have substantiated the effectiveness of the proposed system.