A Sequential Learning Algorithm for RBF Networks with Application to Ship Inverse Control
A sequential learning algorithm for constructing radial basis function (RBF) network is introduced referred to as dynamic orthogonal structure adaptation (DOSA) algorithm. The algorithm learns samples sequentially, with both structure and connecting parameters of network are adjusted on-line. The algorithm is further improved by setting initial hidden units and incorporating weight factors. Based on the improved DOSA algorithm, a direct inverse control strategy is introduced and applied to ship control. Simulation results of ship course control simulation demonstrate the applicability and effectiveness of the improved DOSA algorithm and the RBF network-based inverse control strategy.
KeywordsRadial basis function network Sequential learning Inverse control
Unable to display preview. Download preview PDF.
- 7.Yin, J.C., Dong, F., Wang, N.N.: A Novel Sequential Learning Algorithm for RBF Networks and Its Application to Dynamic System Identification. In: 2006 International Joint Conference on Neural Networks, pp. 827–834. IEEE Press, New York (2006)Google Scholar
- 8.Widrow, B., Wallach, E.: Adaptive Inverse Control. Prentice Hall, Upper Saddle River (1996)Google Scholar
- 11.Chislett, M.S., Strom, J.T.: Planar Motion Mechanism Tests and Full-scale Steering and Manoeuvring Predictions for a Mariner Class Vessel. Technical Report, Hydro and Aerodynamics Laboratory, Denmark (1965)Google Scholar