Research on path tracking controller of mining truck based on improved LQR

To achieve accurate path tracking for unmanned mining trucks when facing roads with a large range of continuous curvature changes, an improved linear quadratic regulator (ILQR) control method combining Fuzzy Logic-based Preview Distance Design (FBPD) and Adaptive Weight Matrix Design trained by Modified Particle Swarm Optimization (MPSO) is brought up in this study. The design of preview distance based on fuzzy logic considers road curvature and truck speed as inputs and the best preview distance as output. The Modified PSO method is adopted in this study to train the weight matrix of LQR under the conditions of different truck velocity intervals and road curvature intervals. The optimal weights under different path curvature intervals and truck velocity intervals are obtained, and finally made into two-dimensional look-up table to realize weight self-adaptation. Finally, the simulation test conducted on the truck whole dynamics model independently developed by Simulink and the real truck verification are used to compare the tracking effect of this algorithm with pure tracking and other algorithms. The algorithm shows good path tracking performance.