出版社：Japan Society for Fuzzy Theory and Intelligent Informatics
摘要：It is very difficult for a conventional PID controller to achieve good control of a small vessel under any operating conditions, because the dynamics of the vessel varies significantly with speed, weight, and disturbances such as wind, waves, tidal currents, etc. Since the autopilots, which have already been developed for large vessels, require the system identification or the reference model, they are not suitable for small vessels. The purpose of this research is to realize the adaptive autopilot for small vessels without the system identification or the reference model, which can be achieved by unsupervised learning. The adaptation needs the adequate evaluation method of the behavior of the system under control. We propose the evaluation method employing Trajectory On Phase diagram (E-TOP) as the new evaluation method for small vessels in order to build up the adaptive autopilot for them. Moreover, the new adaptive autopilot system is examined with a vessel simulator and a real pleasure boat to show the E-TOP method is much more effective rather than a conventional PID controller.
关键词：adaptive autopilot ; PID controller ; fuzzy logic ; unsupervised learning algorithm ; evaluation with trajectory on phase diagram (E-TOP)
