Adaptive neural sliding mode control for two wheel self balancing robot

This paper presents an adaptive sliding mode controller based on a neural network to a control reference trajectory for a two-wheeled self-balancing robot system (TWSBR). In the proposed control scheme, a three-layer neural network is applied to online estimate the unknown model parameters. In addition, a robust adaptive controller is also used to compensate for the estimating errors and uncertainties of the TWSBR control system. The design of the online updating laws for parameters of the neural network and the uncertainties compensator is derived by using the Lyapunov stability theorem. Therefore, the proposed controller can guarantee stability and robustness in the presence of uncertainties. Based on the simulation and experimental results, we found that the output values of the TWSBR control system follow the desired values near a neighborhood of zero, provided evidences to verify the effectiveness and performance of the proposed controller.