A NEURAL NETWORK AUTOPILOT FOR SHIP CONTROL

With the search for the definitive adaptive ship control system inconclusive there is cause to investigate alternative strategies. The natural successor of adaptive control is intelligent control. This paper briefly describes modern adaptive methods and introduces the topic of intelligent control. Intelligent control consists of three methodologies: expert, fuzzy and neural. By comparing the relative merits of each approach, this paper shows that the neural network is the most appropriate method for further investigation into the ship control problem. An implementation is presented and the results of simulation trials are offered as a comparison with traditional control methods. The paper concludes by proposing strategies for further research into track control. Traditional optimal or predictive techniques are considered satisfactory for implementation, however in order to make these algorithms adaptive some intelligence is required. It is proposed to use neural methods to complete the solution.

• Supplemental Notes:
  • Maritime Communications and Control, 3rd Intl Conf; 7-8 July 1993; London, UK. Organised by Inst Marine Engrs, London, UK. Procs. Pubs by Marine Management (Holdings) Ltd, London, UK, ISBN 0-907206-52-2. Ppr 4, p 47 [14 p, 19 ref, 18 fig]
• Authors:
  • Witt, N A
  • Miller, K M
• Publication Date: 1993

Language

• English

Subject/Index Terms

• TRT Terms: Artificial intelligence; Automatic pilot (Aircraft); Automatic pilot (Ships); Control systems
• Subject Areas: Data and Information Technology; Marine Transportation;

Filing Info

• Accession Number: 00707918
• Record Type: Publication
• Source Agency: British Maritime Technology
• Files: TRIS
• Created Date: Aug 14 1995 12:00AM