Motion Control of Mini Underwater Robots Based on Sigmoid Fuzzy Neural Network

Abstract

Aiming at high maneuverability and ability to avoid obstacles in motion control of mini underwater robots, a novel method of control based on sigmoid fuzzy neural network was presented. The structure of fuzzy neural network was constructed according to the moving characters, and the learning algorithm which calculated dynamic learning ratio based on least disturbance was deduced in detail. Finally, simulation and lake experiments were carried out on "WEILONG" mini underwater robot. The results show that dynamic learning ratio keeps the learning of neural network stable and fast, and the operating speed was picked up greatly on the basis that there is no loss for integral control quality. The response ability is improved, which meets the requirement of real-time control.