Tractor path following control usually suffers from the uncertainties and disturbances especially when tractors are faced with uneven soil. As the main actuator of path tracking control, a electro-hydraulic coupling system has parameter uncertainties and disturbances. Conventional methods pay little attention to the electro-hydraulic coupling system which can generate poor path tracking performance. A layered multi-loop robust control architecture (LMLRC) is proposed including a path tracking layer, an angle tracking layer, and a current tracking layer. The path tracking layer is obtained by model predictive controller (MPC) considering uncertainties and disturbances. The angle tracking layer has a sliding mode controller (SMC) which takes parameter uncertainties and disturbances of electro-hydraulic coupling system into consideration. It is designed to follow the target angle generated by path tracking layer and outputs a tracking torque. The current tracking layer is designed by Proportion-Integration-Differentiation (PID) controller to track the torque output by angle tracking layer. To verify the proposed controller, simulation test and hardware in the loop (HIL) test were implemented. The results show that when faced with uncertainties and disturbances the LMLRC is more effective at path following. Moreover, better dynamic stability is exhibited when compared with other controllers.

拖拉机路径跟随控制通常会受到不确定性和干扰的影响，尤其是当拖拉机面临不平整的土壤时。电液耦合系统作为路径跟踪控制的主要执行器，具有参数不确定性和扰动性。传统方法很少关注会产生较差路径跟踪性能的电液耦合系统。提出了一种分层多回路鲁棒控制架构（LMLRC），包括路径跟踪层、角度跟踪层和电流跟踪层。路径跟踪层是通过考虑不确定性和干扰的模型预测控制器（MPC）获得的。角度跟踪层有一个滑模控制器（SMC），它考虑了电液耦合系统的参数不确定性和扰动。它旨在跟踪路径跟踪层生成的目标角度并输出跟踪扭矩。电流跟踪层由比例-积分-微分（PID）控制器设计，通过角度跟踪层跟踪扭矩输出。为了验证所提出的控制器，实施了仿真测试和硬件在环 (HIL) 测试。结果表明，当面临不确定性和干扰时，LMLRC 在路径跟踪方面更有效。此外，与其他控制器相比，表现出更好的动态稳定性。进行了仿真测试和硬件在环（HIL）测试。结果表明，当面临不确定性和干扰时，LMLRC 在路径跟踪方面更有效。此外，与其他控制器相比，表现出更好的动态稳定性。进行了仿真测试和硬件在环（HIL）测试。结果表明，当面临不确定性和干扰时，LMLRC 在路径跟踪方面更有效。此外，与其他控制器相比，表现出更好的动态稳定性。