Development of a path-tracking control system based on model predictive control using infrastructure sensors

This paper describes the development of an infrastructure-based path-tracking control system. The control system consists of an infrastructure sensor module, a vehicle controller and actuator modules. A reference path is defined from a start point to a destination and a modified reference path is generated to obtain a safe vehicle trajectory for collision avoidance in the case where obstacles and other vehicles exist. Receiving information about vehicle position, heading angle and obstacles surrounding the vehicle from an infrastructure sensor module, the vehicle controller calculates control inputs such as steering wheel angle, throttle angle and brake torque to track the modified reference path which guarantees minimum clearance to obstacles. The vehicle controller comprises three parts: a path-generation algorithm, a path-following controller and a speed controller. The path-generation algorithm generates the modified reference path using the model predictive control method. The path-following controller calculates steering angle in order to track the modified reference path. From the speed controller, throttle and brake control inputs are generated to follow the reference velocity. The path-tracking control system was implemented using a test vehicle and an infrastructure sensor module.