Collision Avoidance System for Motorcycles (Automated Steering Control Using Model Predictive Control)

Abstract

This study examines a collision avoidance system for motorcycles by automated steering control. The Model Predictive Control (MPC) is applied to the automated steering control with constraints on the motorcycle rolling motion by considering the rider's body movement in emergency avoidance. Two types of MPC controllers are designed and evaluated by the simple computer simulation of obstacle avoidance. One is a controller with a constraint on the rolling angle. The other is a controller with a constraint on the roll rate. Both controllers improve obstacle avoidance performance in comparison with the Liner Quadratic (LQ) controllers as conventional control theory. The MPC controller with the constraint on the roll rate decreases the roll rate under the permissible rolling angle. The controller with constraints on not only the roll rate but also the steering torque suppresses the large steering torque input. The improved avoidance performance by the MPC controller is observed at other vehicle velocities.