Neural Network-Based Robust Synchronous Computed Torque Control for Redundant Driven Cutting Arm

Developing a superior-performance roadheader position control strategy in tunnel section forming operations is crucial for preventing over-under excavation. However, influenced by complicated external loads and inevitable modeling errors, high-precision trajectory tracking control for roadheaders has become very difficult. Thus, this article designs an adaptive synchronous computed torque controller for the roadheader’ s hydraulic drive system. Firstly, a complete cutting system dynamic model is constructed using the Lagrangian method, and the optimization of cylinder driving force is achieved using the minimum 2-norm method. Then, combined with the idea of synchronous control, the synchronization error and coupling error of the cylinder are defined. Further, a synchronous computed torque controller in joint space is designed to avoid internal force coupling caused by asynchronous drive cylinders. By introducing neural networks and robust integral terms, adaptive compensation for the estimation term of uncertainty disturbances is effectively achieved. The asymptotic stability of the cutting system is proved using Lyapunov stability theory. Finally, the proposed controller's effectiveness is verified through comparative experiments on a roadheader prototype with redundant drives and uncertain disturbances.