Fuzzy adaptive control for nonlinear systems. Real-time implementation for a robot wrist

In this paper, an adaptive Takagi-Sugeno fuzzy controller for continuous nonlinear systems is proposed. The controller is designed under the constraint that only the output of the plant is measured. A state observer is thus introduced in the global control structure. The nonlinear plant is approximated with a Takagi-Sugeno fuzzy system whose parameters are adjusted via adaptive laws. Based on this representation, the control law is designed according to the Lyapunov and passivity theories. Stability of the algorithm is established with tracking errors converging to zero. Finally, in order to illustrate the feasibility of the proposed methodology, a real-time implementation for controlling a robot wrist is presented.