Hybrid neuro-fuzzy network identification for autonomous underwater vehicles

Abstract

Autonomous Underwater Vehicles (AUVs) are ideal platforms for aquatic search and rescue operations and exploration. The AUV poses serious challenges due to its complex, inherently nonlinear and time-varying dynamics. In addition, its hydrodynamic coefficients are difficult to model accurately because of their variations under different navigational conditions and manoeuvring in uncertain environments. This paper introduces an identifier scheme for identification of non-linear systems with disturbances based on Hybrid Neuro-Fuzzy Network (HNFN) technique. The method comprises of an automatic structure-generating phase using entropy based technique. The accuracy of the model is suitably controlled using the entropy measure. To improve the accuracy and also for generalization of the model to handle different data sets, Differential Evolution technique (DE) is employed. Finally, Hardware In-Loop (HIL) simulation and real-time experiments using the proposed algorithm to identify the 6-DOF UNSW Canberra AUV's dynamics are implemented. The modelling performance and generalisation capability are seen to be superior with our method. © 2013 Springer International Publishing.