Neural network and PID hybrid adaptive control for horizontal control of shearer

The shearer is a most important tool for coal mining. The task of horizontal control of shearer is to make drum track the coal-rock interface accurately in order to improve recovery. A neural network adaptive controller combined with PID is presented considering nonlinear and uncertainty of the electro-hydraulic servo system used for horizontal control. The neural network is trained on-line with the reference signal and the feedback error. Feedback-error-learning scheme is employed to keep the stability of system as well as a short learning time. Simulation and experiment on the hybrid of neural network and PID controller are presented. The results show satisfactory performance of the controller.