Tuning of Neuro-fuzzy Controller by Real-Coded Genetic Algorithms with Application to an Autonomous Underwater Vehicle Control System

This paper proposes a neural-fuzzy controller (referred to as NFLC) tuned automatically by genetic algorithms (GA). A real-code method is used to encode the GA chromosome, which consists of the width and center of the membership functions, and the rule sets of the controller. Dynamic crossover and mutation probabilistic rates are applied for faster convergence of the GA evolution. Application of the NFLC to an autonomous underwater vehicle (AUV) is investigated. The NFLC shows considerable robustness and advantages compared with a manually-tuned conventional fuzzy logic controller which are applied to the same AUV.