Tube-Based Model Predictive Control for Unmanned Vehicles Trajectory Tracking

With the advancement of autonomous driving technology and the accelerated commercialization of mobile robots, real-time trajectory tracking and control of unmanned vehicles in complex environments have become critical challenges to system reliability. To better address the insufficient robustness and limited tracking accuracy of traditional Model Predictive Control (MPC) algorithms in local trajectory tracking for unmanned vehicles, this paper proposes a Robust Tube-based Model Predictive Control method. First, a vehicle system model is established and the nonlinear system is linearized. Subsequently, by constructing an error model between the nominal system and the actual system, vehicle constraints are integrated to formulate a robust objective function. The feedback gain is then solved through the Linear Matrix Inequality (LMI) approach to reduce lateral position errors. Finally, co-simulations using CarsimSimulink are conducted to evaluate the improved MPC algorithm under various conditions. Experimental results demonstrate that the proposed Tube-based MPC method achieves superior tracking accuracy and exhibits enhanced robustness compared to conventional approaches.