Research on Obstacle Avoidance Tracking Control Method for Unmanned Vehicles Based on Model Predictive Control in Variable Time Domain

Aiming at the obstacle avoidance tracking control problem of unmanned vehicles, a model predictive controller is designed in this paper. The controller adopts a layered structure. The upper layer performs local path planning during obstacle avoidance based on the nonlinear vehicle point-mass model, and the lower layer implements path tracking based on the monorail vehicle dynamics model. Through the joint simulation of Carsim and Matlab, the genetic algorithm is used to optimize the controller parameters, and the performance of the controller is tested at different speeds. The simulation results show that: compared with the controller before improvement, the controller is at 30km/h, 60km/h and 80km/h, the lateral deviation after obstacle avoidance decreased by 34.8%, 47.2%, and 13% respectively, and the yaw angle deviation decreased by 18.7%, 49.9%, and 29% respectively, Finally, smooth obstacle avoidance at various speeds is achieved through layered control.