This paper addresses the impact of complex mining environments and the nonlinear dynamics of hinged mining trucks on reverse path tracking control for autonomous mining trucks. We propose a neural-network-based sliding mode control (NN-SMC)-based control strategy for reverse motion to improve tracking accuracy and robustness. First, a tractor–trailer dynamic model is built, and the force characteristics at the coupling joint are analyzed to derive the reverse interaction forces, which simplifies trailer modeling and avoids the influence of uncertain tractor parameters. Next, a control scheme matching the simplified model is developed, where an optimized sliding surface is designed and a neural network adaptively tunes control parameters to reduce chattering and improve adaptability to challenging conditions. Finally, hardware-in-the-loop tests validate the simulation results. Both simulation and experiments show that, compared with conventional SMC, the proposed method reduces lateral displacement error by 13.98% and heading error by 18.96%, demonstrating the effectiveness of the control approach.
Yang et al. (Tue,) studied this question.