Type-2 fuzzy logic neural network control and target following for Remote Operated Vehicles

Accurate control and target following are very important for Remote Operated Vehicles subsea engineering. In order to make Remote Operated Vehicles task more effective and accurate, type-2 fuzzy logic neural network has been introduced to model and minimize the effects of uncertainties in rule-base fuzzy logic system. In the online learning strategy, the gradient descent method has been used for online training. The results of tank experiments have proved that the controller can improve the computation efficiency, reduce control errors, vibration and overshoot. Thus the controller displays strong robustness in the underwater robotic control. In the 3-D target following simulations, ROV can precisely follow the target in still and current underwater enviroment. These simulations verify the controller´s capacity to realize 3D-trajectory control and target following. It can realize 3D-target following curve tracking, obstacle avoidance and guarantee underwater task accomplishment.