Evolutionary Algorithms Based Model Predictive Control for Vehicle Lateral and Roll Motion Control

This paper proposes a trajectory tracking strategy based on a linear extended bicycle vehicle model and a linear tire model. Model predictive control is used to track the lateral positions, sideslip angle, yaw rate, and load transfer ratio. To solve the MPC objective function in the field of vehicle stability, evolutionary algorithms are implemented, including multi-objective genetic algorithm (MOGA), particle swarm optimization (PSO), and hybrid particle swarm optimization genetic algorithm (HPSOGA). These algorithms are used to optimize the weights as well as the control input. To handle constraints violations, the penalty function method is exploited. To demonstrate the sensitivity to the respective parameters, a local sensitivity analysis is performed on the sideslip angle. The obtained results show a simultaneous change in all output variables due to the state coupling effect. A root mean square error (RMSE) of hybrid algorithm for sideslip angle is calculated as 0.032 and for PID controller is obtained as 0.18. It is found that HPSOGA (hybrid algorithm) based MPC performs better in comparison to PID controller.