Adaptive neural network based PI sliding mode control of nonholonomic mobile robot

In this paper, A hybrid control algorithm which combines classic backstepping kinematic control(CBKC) and PI-type sliding mode control (SMC) based on both radial basis function neural network (RBFNN) and saturation function (SF) is proposed for trajectory tracking of nonholonomic mobile robot. For the stability and robustness, PI-type sliding mode control maintains the effectiveness of overcoming the uncertainties of mobile robot. Because of the existence of discontinuity and chattering problem when exclusively using PI-type sliding mode control, a saturation function and RBFNN are cooperatively utilized to alleviate the problems and enhance the adaptability for random uncertainty. The simulation results demonstrate that the controller has smooth control input, and the tracking errors of nonholonomic mobile robot are eliminated.