Neural-Network-Based T-S Fuzzy Control Of Robot Fish

In the dissertation, to research on motion control and motion simulation of robot fish is the main task , based on the "FangSheng - I " robo-fish of Harbin Engineering University AUV Lab.Fish and such swimming animal have excellent propulsion and maneuvering system, whose advantages are high efficiency and low noise. Making use of the biomimetic technology, We have developed new system imitated dolphin or tuna.Thesis is make up of two parts. One part is to construct a motion model of 5-direction freedom, which is take advantage of the result that is analyze numerically the unsteady hydrodynamic force of the machine fish by surface panel method. The other part is design a controller for basic motion of machine fish. The controller is in view of Neural-Network-Based T-S Fuzzy control method. On this condition, we debugged the controller. As a result, it is identified that the Neural-Network-Based T-S Fuzzy control method is right and feasible.Based on the results of robot fish' s motion simulation, the simulation experiment error and the reasons were discussed and analyzed in this dissertation. On all accounts, it is practicable for the future research on robot fish.