Research On Adaptive Neural Network Fuzzy Control Strategy In The Inverted Pendulum

The inverted pendulum is a kind of typically nonlinear, multi-variable, unstable, strong-coupling and fast system. The research of inverted pendulum can come down to the research of the multi-variable and absolutely unstable system, so its control methods and ideas have comprehensive purposes to the general complex industry process.As an emphasis, this paper proposes a novel intelligent control strategy-ANFIS (adaptive neural network-based fuzzy inference system) to model and control the complex inverted pendulum. This strategy adopts the merit of Sugeno fuzzy model that can express dynamic characteristic of nonlinear systemeasily. ANFIS strategy does not need to vary structure of whole system, so it is simple to realize only by optimizing control parameters meanwhile using BP-GDA hybrid algorithm. Based on captured objective input-output data couple, a new ANFIS framework was used to construct fuzzy membership function and fuzzy rules. Then we will control one and two stages inverted pendulum effectively depend on modified Sugeno fuzzy model.Control experiments show that, comparing with PID and traditional fuzzy control strategies, this ANFIS strategy avoids drawback of designing fuzzy membership function and fuzzy rules subjectively and the control performance is satisfied, therefore, it points out a new method to resolve identification and control problem of unknown complicated nonlinear system.