Intelligent ship autopilots––A historical perspective

The paper presents a historical perspective of the development and use of the so-called intelligent paradigms of fuzzy logic and/or neural networks in ship autopilot designs. After a brief review of the development of the first PID autopilots, the paper describes how early work using fuzzy logic techniques to describe and model the actions of helmsmen led to the development of the first fuzzy autopilots and how parallel advances in self-organising and adaptive network fuzzy inference systems were utilised to refine the early fuzzy autopilot designs. The use of artificial neural networks in adaptive and model reference approaches is also highlighted as is their use in optimising the parameters of neurofuzzy designs. The paper concludes with some observations on stability issues and suggestions for future developments in intelligent ship autopilots. [Copyright &y& Elsevier]