Self-adaptive neuro-fuzzy systems for autonomous underwater vehicle control

Abstract

This paper presents a systematic approach for developing a concise self-adaptive neurofuzzy inference system (SANFIS) with a fast hybrid parameter learning algorithm for on-line learning of the control knowledge for autonomous underwater vehicle (AUV) control. The multi-layered network structure of SANFIS incorporates fuzzy basis functions for better function approximations. We investigate three SANFIS structures with three different types of fuzzy IF-THEN rule-based models and cast the rule formation problem as a clustering problem. A recursive least-squares algorithm and a modified Levenberg-Marquardt algorithm with limited memory are exploited to accelerate the parameter learning process. This hybrid parameter learning algorithm together with an on-line clustering technique and rule examination provide SANFIS with the capability of selforganizing and self-adapting its internal structure (i.e. the fuzzy rules and term sets) for learning the required control knowledge for an AUV to follow desired trajectories. Computer simulations for modeling a control system for an AUV have been conducted to validate the effectiveness of the proposed SANFIS.