Radial basis function networks for ship steering control

Unar, M.A. and Murray-Smith, D.J. (1997) Radial basis function networks for ship steering control. In: 12th International Conference on Systems Engineering (ICSE '97), Coventry, UK, 9-11 Sep 1997, pp. 700-705.

Abstract

This paper investigates the application of radial basis function (RBF) networks for ship steering control systems. For training and testing purposes, models of three different ships are considered. A significant feature of the models is that their parameters change with forward speed of the ships. A supervised learning strategy is applied for the training of the networks and PD/PID controllers designed at different forward speeds are used as supervisors of the networks. The networks are then capable of yielding satisfactory performance at varying forward speeds within the range considered in the training phase. The paper also includes a comparison of the performance of the networks with that of networks trained by using the back propagation rule for the same application.

• Item Type: Conference Proceedings
• Additional Information: Proceedings ISBN: 0905949633
• Keywords: Radial basis function networks, multilayer perceptrons, ship autopilots, supervised learning
• Status: Published
• Refereed: Yes
• Glasgow Author(s) Enlighten ID: Murray-Smith, Professor David
• Authors: Unar, M.A., and Murray-Smith, D.J.
• Subjects: Q Science > QA Mathematics > QA76 Computer software; T Technology > TJ Mechanical engineering and machinery; T Technology > TK Electrical engineering. Electronics Nuclear engineering; V Naval Science > VM Naval architecture. Shipbuilding. Marine engineering
• College/School: College of Science and Engineering > School of Engineering > Systems Power and Energy