Influence of Preview Distance on LMPC-Based Path Tracking

In previous research, it has been found that linear model predictive control (LMPC) has advantages in dealing with system constraints, while preview control can introduce nonlinear feedforward information into path tracking. But so far, the advantages of these two control methods have not been combined. For the combination of LMPC and preview control, the most important problem to be solved is how to set the preview distance. To solve this problem, we added different preview distances for the LMPC-based path tracking controller under different working conditions. Through the simulation results, we reached five conclusions. Firstly, the performance of LMPC can be improved by adding the preview distance. Secondly, for the LMPC-based path tracking controller, there is an optimal preview distance. Thirdly, the optimal preview distance is related to the longitudinal velocity, and usually, a longer preview distance should be taken when the longitudinal velocity is higher. Fourthly, the correlation between the curvature of the reference path and the optimal preview distance is weak. Fifthly, adjusting the preview distance does not affect the real-time performance of the LMPC controller.