System Identification of an Inverted Pendulum Using Adaptive Neural Fuzzy Inference System

The objective of this paper is to illustrate the efficiency of adaptive neural fuzzy inference system (ANFIS) in identifying a nonlinear single-input multiple-output (SIMO) system. The SIMO system used for demonstration is cart-inverted pendulum, which is well known for its highly nonlinear, unstable, and under-actuated nature. The ANFIS model of cart-inverted pendulum (CIP) is designed in Matlab Simulink environment using input-output data obtained from nonlinear mathematical model. The simulation responses for different initial conditions are obtained from ANFIS model which are further compared to the mathematical model of the system. It was observed that within the trained operating range, ANFIS model exactly replicated the nonlinear mathematical model of the system while a little deviation is observed outside the trained operating range. Thus, the authors propose to use ANFIS for system identification from experimental input-output data when the system parameters are unknown or uncertain.