The Research On A Self-Adapting Neural Network Control Method Of Dynamic Positioning To Line For Underwater Vehicle

System modeling, simulating and controlling are studied in this paper, based on the project 'The Equipment Used by ShengLi Petrol Field for Detecting and Repairing Pipes and Cables under Seabed of The Shallow Sea' . Due to its work feature, the underwater vehicle studied here should be positioned to line accurately. As one of the dynamic positioning methods, positioning to line is different from the fix point positioning as normally mentioned, and needs new control strategy.Considering the underwater vehicle itself is a system which is nonlinear and strong coupled and the hydrodynamic parameters is difficult to get, which increases the complexity of its control. For such a kind of system, neural network control proved to be an efficient method, and have been already applied in the control of the underwater vehicles and the underwater robots. The neural network has ability to approach arbitrary nonlinear function, and structure and algorithm of studying are simple and clear. By the study of the neural network, Can find some P, I, D parameter under optimum controlling rate.This paper divide the system into three subsystems and discussed the underwater vehicle' s lengthways, horizontal and heading freedom degrees and their controlling methods respectively. Considering that the exterior interference mainly comes from the sea flow when it is underwater, this paper prove the adaptive capacity of neural network control method against the model uncertainties and external disturbances, such as the variations of the sea current' s speed and direction, and it also compared itself with the PID controller in common use. Finally, from the simulating results, we can see that this...