An Investigation of Classical Model Predictive Controller Path Tracking Performance of a Two-Wheel and Four-Wheel Steering Vehicle

Studies on self-driving vehicles have become a trend in recent years, and many systems have been developed to enable autonomous manoeuvre. Various methods have been used to improve path-tracking algorithms which increase vehicle performance, including tracking accuracy and stability. Path tracking is one of the primary problems for autonomous vehicles where the vehicle deviates from target paths, which leads to unnecessary counter-correction. Conventional front wheel steering is unable to satisfy the manoeuvre with a high lateral acceleration since the front steering angle is limited in accurately responding to vehicle dynamics. Moreover, the characteristics of front wheel steering vehicles affect handling stability due to the fact that the turning radius is larger than the vehicle itself. This disadvantageous can compromise safety during under-steer and over-steer situations. The main objective of this preliminary study is to investigate the performance of a four-wheel steering system (4WS) in path tracking for autonomous vehicles using a classical model predictive controller (MPC). Conventional two-wheel steering (2WS) tracking performance following the desired driving system with the same MPC controller is compared with 4WS vehicles. The driving system is developed using Driving Scenario Designer to extract the desired yaw angle and lateral position for controller references constructed in Matlab Simulink. Fixed MPC constraint, prediction, control horizon, yaw angle and lateral position weights were used to compare the performance between 2WS and 4WS vehicles. The simulation results show that 4WS is three times better than 2WS vehicles in tracking predetermined paths. 4WS vehicle show 74.55% better performance in lateral position tracking and 68.75% better performance in trailing predetermined yaw angle value. The simulation data from the preliminary study will be used as a guideline to develop an advanced controller of 4WS vehicles.