Variable Tube-Based Model Predictive Trajectory Tracking Control Strategy with Adaptive Load for Automatic Truck

High-precision trajectory tracking is the basis for achieving safe driving of automatic trucks. However, variable loads can cause changes in vehicle parameters, leading to a severe decline in the tracking accuracy of autonomous trucks. To address this problem, we propose a variable Tube-based model predictive trajectory tracking control strategy with strong robustness to load changes. The variable Tube-based model predictive trajectory tracking control strategy includes the radial basis function (RBF)-Ridge observer and dynamic cross-section Tube-based model predictive control (DCTMPC). The data-driven RBF-Ridge observer performs real-time estimation of parameters affected by load to reduce the range of parameters disturbance. The RBF-Ridge observer calculates the optimal ridge parameter based on sensor signals, eliminating the impact of multi-collinearity and unbalanced on parameter estimation accuracy. The DCTMPC adaptively reconstructs the model and adjusts the state Tube based on observer estimation to improve the path tracking robustness of automatic trucks. Hardware-in-the-loop tests show that the proposed strategy can accurately calculate the truck parameters affected by load changes and improve the truck trajectory tracking accuracy under different loads.