Robust neuro-fuzzy model-following control of robot manipulators

A robust neuro-fuzzy model-following control system is proposed for robot control with torque disturbance and measurement noise. The control objective is obtained by tailoring a nominal adaptation process of weights and a fine tuning mechanism to overcome the equivalent uncertainty. The major difference comparing with previous approaches is that a novel fuzzy system is introduced such that the fuzzy rules are in the form of "IF situation THEN the control input" rather than "IF situation THEN the value of some nonlinear functions". Using Lyapunov stability method, the uniform ultimate boundedness of tracking error has been proved.