Study On The Recurrent Neural Network Dynamical Identification And Its Application To Ship Motion Control

Along with the development of shipping business, ships are becoming bigger, faster and more intelligent thus better performance of maneuver is demanded. To ensure safety and improve economy, it is necessary to adopt new control theories and techniques, and to research for better control strategies. The application of neural network identification techniques in ship control became an important research area in recent years. This thesis concentrates on studying the identification techniques of network and two novel learning algorithms for system identification are proposed. This thesis also discussed the applications of network identification in ship motion control.Elman recurrent neural network has been widely used in time-varying system identification applications. By combining radial basis function network with Elman network, we got RBF-Elman network. Further improvements have been made by adding the delayed output message in the input layer, resulting network is reffered to as Output-Feedback-Based Elman network. The network is capable of identifying time-varying dynamics, its learning speed is improved by the linear connection of hidden layer and output layer. Its satisfactory identification capability and adaptive capability are denmonstrated by time-varying system identification experiment.To meet the demand of on-line application of neural networks in control system, a novel sequential learning algorithm of RBF network is proposed referred to as dynamic tracking model selection (DTMS) algorithm. The self-adaptation ability of the resulting network was demonstrated in identification of systems with static and time-varying dynamics. The algorithm is also featured by small mumber of tuning parameters, explicit meaning of parameters, adaptive adjustment of parameters and its robustness to changes of parameters.Aiming at the nonlinear and time varying characteristics of ship motion, also for application of control strategy, the network-identification-based predictive control strategy is proposed. The strategy employs neural networks for on-line sequential learning, and tracks the changes of ship motion dynamics efficiently. Finally, the proposed strategy was applied in ship course tracking control simulation and the satisfying performances demonstrate the feasibility and effectiveness of the ship control strategy.