Robust adaptive control of robots using neural network and sliding mode control

This paper presents a method for designing robust adaptive control of strict-feedback systems with function uncertainties and disturbances. A backstepping-based neural network controller is connected in parallel with a sliding mode controller to utilize best advantages of two approaches. The neural network is used to approximate the uncertainty functions, where the weighting coefficients of the neural network are trained online. The robust adaptive control law is designed based on control Lyapunov function by using backstepping techniques and sliding mode control, thus global asymptotic stability is guaranteed for the case of ideal implementation of the neural network. The proposed controller is applied to an n-degrees-of-freedom robot. The simulation results demonstrate the effectiveness of the proposed method.