Research on fuzzy neural network-based sliding mode balance control of Acrobot

Abstract

Acrobot is a kind of under-actuated two-link rods robot. The control methods of Acrobot are complicated, but doing research on it is valuable to the applications and the study of nonlinear systems. Balance control and swing-up control are two main control areas of Acrobot. This paper centers on its balance control and combines the advantages of sliding mode control and fuzzy neural network to control Acrobot system. The simulations show that the new control strategy proposed in this paper is more effective than SMC (Sliding Mode Control) or ANFIS (Adaptive-Neural Network-based Fuzzy Inference System) respectively used to control Acrobot.