Adaptive neural sliding mode control of an uncertain permanent magnet linear motor system with unknown input backlash in laser processing

In this article, a novel robust and high-precision control scheme is developed for a permanent magnet linear motor system in large-range laser processing subject to unknown system uncertainties and input backlash. First, to enhance the robustness of the controller, an adaptive neural sliding mode control (SMC) scheme is proposed. Besides, to address the system uncertainties, a novel simplified radial basis neural network (RBFNN) is applied in the adaptive SMC scheme. Subsequently, considering that the unexpected chattering introduced by SMC strategy, another RBFNN is employed to approximate the optimal switching gain. In addition, considering that the nonlinear input backlash caused by actuator can also cause the fluctuation of the input signal, an inverse backlash model with the method of reverse compensation is proposed to further minimize the effect of the input backlash on control accuracy. Finally, simulation results, validation experiments, and laser processing experiments are provided to demonstrate that the proposed control scheme significantly reduces the impact of chattering and input backlash. Additionally, it exhibits outstanding control accuracy.