Adaptive robust backstepping control based on radial basis neural network for linear motor drives

This work presents an adaptive backstepping controller using a radial basis function neural network (RBF-NN) for position control of a linear motor drive with parameter uncertainties, discontinuous friction and unknown external disturbances. Initially, a robust control scheme is developed to ensure asymptotic stability. To avoid conservative tracking performance, we propose an adaptive robust backstepping law incorporating an RBF-NN to estimate lumped uncertainties and disturbances. The dynamic determination of the approximation error upper bound eliminates discontinuities in the adaptive control law. The RBF-NN's characteristics are utilised to establish the existence of solutions for the system, ensuring that the adaptive control law satisfies the Lipschitz continuity condition. The developed scheme ensures global asymptotic stability under bounded disturbances. Simulation results validate the proposed scheme's effectiveness in achieving precise positioning and reducing chattering compared to a robust backstepping controller, a fast nonsingular terminal sliding mode controller and an adaptive recursive terminal sliding mode controller.