Adaptive fuzzy-Chebyshev network control of a conducting polymer actuator

In this paper, the tip displacement of a conducting polymer actuator (CPA), which has an unknown and time-varying dynamics, is controlled using three different model-free adaptive controllers. First, conventional indirect adaptive fuzzy and Chebyshev functional-link network controllers are introduced. Second, a novel hybrid network which combines the capability of conventional fuzzy system and Chebyshev functional-link network is proposed for online identification and control of the CPA. The stability of the proposed controllers and boundedness of the closed-loop signals are proven using Lyapunov stability approach. The designed adaptive controllers are implemented experimentally in real-time. The performances of the controllers in tracking a square and a sine reference are compared in terms of root-mean-squared error (RMSE), required input signal power (ICVP) and settling time (t(s)) performances.