Trajectory Tracking Control of Unmanned Vehicles Based on Linear Time-Varying Model Predictive Control

In order to solve the problem of large errors in the track tracking control of unmanned vehicles during running, a linear time-varying model predictive control (LTVMPC) device was designed, and the output and track error constraints were introduced. In order to avoid hard constraints leading to no linear solution in the rolling optimization process, relaxation factor vector extension was introduced into the objective function to solve the feasible domain, and then the LTVMPC optimization problem was transformed into an online quadratic programming problem. Aiming at the problem that trajectory tracking control is difficult to apply to the obstacle environment, combining with reinforcement learning algorithm, a trajectory planning layer is established on the trajectory tracking layer, and a double-layer trajectory tracking controller with autonomous obstacle avoidance function is designed to realize the autonomous obstacle avoidance function of vehicles in the course of trajectory tracking. Finally, Matlab/Simulink and vehicle dynamics simulation software Carsim were used to design the co-simulation. The results show that the method can adapt to the tracking of double-shift track under different speed conditions, and the tracking accuracy and vehicle stability are significantly improved.