Practical results from full scale sea trials for a self-organising design of fuzzy logic autopilot

Current marine applications are constantly demanding increased performance in the minimisation of both course heading error and rudder activity. Energy efficiency is now also a prerequisite when considering the potential of a novel design of controller for this field of application. Previous studies have established that the implementation of a design of marine autopilot using fuzzy logic control can provide significant performance advantages when compared to the conventional proportional plus integral plus derivative (PID) alternative. The emphasis of this work concerns the control for small marine vessels. The paper discusses the implementation of both PID and fuzzy logic designs of autopilot in full scale sea trials. Highlighted are the significant improvements obtained in the vessel performance when undertaking the two modes of operation, these being course changing and course keeping. As an extension of this work the fuzzy autopilot can be modified to become self-organising. Learning about the operating environment may then be achieved to enforce internal modification within the controller's structure and to therefore further improve the levels of performance obtainable. Additional sea trial results for the self-organising version are also analysed both in comparison to the original PID autopilot, and to the fuzzy logic design. Consideration of these presented results allows the derivation of a strategy for the advancement of the technology for the small vessel application.