Research On The Underwater Manipulator Control Method Based On The Neural Network

The research on how to acquire stable movement of the underwater manipulator is crucial when we need to complete the assigned task .The motion control is one of the key technologies in the underwater manipulator's research fields.Taking into account the economic, security and other factors, the research on how to design and debug the controller using the simulation system is great significant because underwater manipulator's working environment is complexity and unknown. In this paper, we discussed the underwater manipulator's dynamics model and motion control issues by investigating two joints underwater manipulator.Because the underwater manipulator is a strong nonlinear and coupling system, the authors used a conventional PID controller, a CMAC and the PID composite controller to simulate the system on computer. Finally the simulation result was given.Therefore, the main content of this paper can be divided into two parts:(1)The research on the modeling processes for an underwater manipulator simulation systems;(2)The research on the underwater manipulator movement control.This paper firstly introduced correlative mathematical knowledge used to derive dynamic model of the underwater manipulator. In the first part of this paper the author mainly introduces how to use Newton-Euler's law to derive the manipulator's dynamic model. Following the author derived the underwater manipulator's dynamics model with an evaluation of the simple hydrodynamic effects. In the second part, the authors introduced the artificial neural network's basic principles. Then the various structures of the NN are analyzed. Following the author introduced the CMAC controller's principle and characteristics. Finally the author designed the CMAC controller and simulated the system by two different methods. The conclusion was given after the results was compared .