Research on model predictive trajectory tracking control based on particle swarm optimization

Abstract This paper focuses on the challenges of low trajectory tracking accuracy in intelligent vehicles caused by different road adhesion coefficients and vehicle speed conditions, aiming to address these challenges with an improved Model Predictive Control (MPC) algorithm. Firstly, a model encompassing vehicle dynamics featuring two degrees of freedom has been formulated, along with a corresponding model to track trajectory errors. Secondly, a Model Predictive Control (MPC) algorithm is designed with trajectory tracking accuracy and control increment as the objective functions. Additionally, incorporating a Particle Swarm Optimization (PSO) algorithm with MPC allows for the dynamic computation of the optimal prediction horizon size. Afterward, a unified simulation model is assembled, employing both CarSim and MATLAB/Simulink, to conduct dual-lane trajectory tracking simulation analysis under diverse road adhesion and vehicle speed conditions. The efficacy of the proposed control algorithms is validated through this process.