Integrated sliding mode control of robot manipulator based on fuzzy adaptive RBF

To solve the uncertainty of the parameters of the manipulator dynamics model, the control accuracy and convergence rate of the system affected by the joint friction and external interference, a compound control strategy based on the manipulator dynamics model is proposed. Firstly, a modified power-of-two convergence law is used and combined with an integral sliding mode to design a sliding mode control term to shorten the convergence of the tracking error. Secondly, the approximations of the uncertain variables of the dynamical model are accomplished by using the three sets of RBF neural networks and introducing an adaptive mechanism for online self-tuning of weights, the approximation errors of the RBF neural networks are compensated by using the sliding-mode control term designed in the previous section. Finally, the fuzzy controllers are utilized to calculate the coupled joint friction and outside disturbances. The simulation works show that comparing with the chunked RBF neural network to approximate the sliding mode control logy, the proposed hybrid control theory reduces the mechanical arm joint angular rate response time by 39.4%, the largest solid-state error was cut by 76.8%, and the medium-sized solid-state error was cut by 62.7%, improved control preciseness and the responsiveness of the spatial trajectory tracking of the manipulator arm's joints.