Modelling and motion control of an autonomous underwater robot using a fuzzy and fuzzy-neural approach

The design of an autopilot for the control of an Autonomous Underwater Vehicle (AUV) is of interest both from the point of view of motion stabilisation as well as manoeuvring performance. Such vehicles are commonly classified as being highly non-linear uncertain systems and are consequently difficult to control effectively using model-based control methods. The modelling task of these vehicles which consists of three-dimensional equations of motion for its hydrodynamical shape is very complicated. The paper describes a modelling, mission and motion control system for an AW based on fuzzy and fuzzy neural techniques. The idea for the fuzzy modelling is an on-line supervisory scheduling system, which chooses the linear function that describes the system. Fuzzy mission control and fuzzy neuro motion control strategies, with the ability to adapt membership functions, extract new rules and forget unused rules, are proposed.