Adaptive model predictive control for intelligent vehicle trajectory tracking based on improved sparrow search algorithm

Abstract Addressing the challenge of balancing control accuracy and computational efficiency for dynamic systems in traditional model predictive control (MPC) due to fixed temporal parameters, this paper proposes an adaptive model predictive controller based on an improved sparrow search algorithm (ISSA) utilizing L-T chaotic mappings. First, the sparrow population distribution is initialized through a combination of the Logistic-Tent chaotic mapping and inverse sine transformation; Second, a momentum-elite guided explorer update mechanism and a follower update strategy integrating multi-learning with Cauchy perturbations are designed. Subsequently, an adaptive fitness function incorporating tracking error objectives is established with Np and Nc as optimization variables. Finally, the ISSA-MPC controller is formed based on an improved SSA algorithm, enabling dynamic parameter adjustment. Hardware-in-the-Loop(HIL) test results demonstrate that compared to traditional fixed-parameter MPC and MPC optimized by particle swarm optimization (PSO), the proposed ISSA-MPC controller effectively reduces lateral tracking error (by at least 5.6%) and front wheel steering angle (by at least 4.8%) across multiple vehicle speeds. This enhances tracking precision, stability, and adaptive capability under varying operating conditions.