A model-based neurocontrol approach for car-following collision prevention

This paper presents a model-based neurocontrol approach for car-following collision prevention systems for Intelligent Vehicle Highway Systems (IVHSs). The controller is synthesized using resolved-acceleration-like control popular in robotics, and an online adaptive neural module. A nominal mass of the vehicle is used as the only dynamic model information in the control. Neural module is designed to adaptively compensate for dynamic model discrepancies, and coupling effects due to lateral and yaw motions. Several simulation test results in the face of different driving conditions are presented to validate the controller.