Fuzzy Control of Autonomous Intelligent Vehicles for Collision Avoidance Using Integrated Dynamics

This study aims to take the first step in bridging the gap between vehicle dynamics systems and autonomous control strategies research. More specifically, a nested method is employed to evaluate the collision avoidance ability of autonomous vehicles in the primary design stage theoretically based on both dynamics and control parameters. An integrated model is derived from a half car mathematical model in the lateral direction, consisting of two degrees of freedom, lateral deviation and yaw angle, with a traction mathematical model in the longitudinal direction, consisting of two degrees of freedom, the longitudinal velocity and rolling velocity of the wheel. The integrated model uses a mathematical power train model to generate the torque on the wheel and connects the two systems via the magic formula tyre model to represent the tyre non-linearity during augmented longitudinal and lateral dynamic attitudes. These mathematical models are represented using MATLAB in the time domain. Fuzzy logic is used for a path-following model to control the vehicle in the longitudinal and lateral directions. The dynamic behaviour is subjectively evaluated using an ISO 3888 test track. The vehicle dynamic response includes the vehicle's path, steering angle, lateral acceleration, yaw rate and longitudinal velocity. The results demonstrate that the vehicle successfully selected the vehicle path within the track limits and avoided the obstacles along its path. These results highlight the importance of implementing both vehicle dynamic systems and autonomous control strategies in a meaningful integrated model for proper vehicle performance testing.