Improved Fuzzy Neural Modeling Based on Differential Evolution for Underwater Vehicles

Autonomous Underwater Vehicles (AUVs) have gained importance over the years as specialized tools for performing various underwater missions in military and civilian operations. This study presents the on-line system identification of AUV dynamics to obtain the coupled nonlinear dynamic model of AUV. This proposed model has an input-output relationship based upon neural fuzzy network (NFN) model technique to overcome the uncertain external disturbance and the difficulties of modelling the hydrodynamic forces of the AUVs instead of using the mathematical model with hydrodynamic parameters estimation. Initially the models’ parameters are generated randomly and tuned by Differential Evolution algorithm (DE) with a set of real inputoutput data. Secondly, the back propagation algorithm based upon the error between the identified model and the actual outputs of the plant are used to adopt the model’s parameters. The proposed NFN model adopts a functional link neural network (FLNN) as the consequent part of the fuzzy rules. Thus, the consequent part of the NFN model is a nonlinear combination of input variables. Simulation results show the superiority of the proposed adaptive neural fuzzy network (ANFN) model in tracking of the behavior of the AUV .