Control Of Steward Platform With PIDNN

The parallel robot has become a hot spot for a long time because of its simple structure, high stiffness, high precision, and low movement inertia. It especially adapted to the task which required high precision, heavy load and limited space. But the control method of parallel robot was principlal problem because of its serious nonlinear and coupling, uncertain mathematical model and complicated control demand. Almost all the traditional control method failed to get a satisfying performance on the parallel robot, the appearing of new computer and the artificial intelligence accelerate the robot controlling system to develop quickly.Firstly, This thesis prents the research status of the parallel robot and the developing status of the computer control system.Secondly, the machine feature of the parallel robot was analyzed, steward trajectory planning methods were studied, This research gave out some planning trace by Matlab, via the math model of electro-hydraulic control system. Then, traditional PID and neural network control theory were introduced. steward tracking control system was simulated, and sum up the traditional PID control and neural control rules, analyse the variable parameters effect the contorl system. The software and hardware of steward motion tracking experimental system wre also introduced.Finally, on the basis of introducing the theory of neural network and PID as a whole, this paper presented a new mixed control method, which was called PID neural network(PIDNN). The PIDNN method combined the strong self-study ability of neural network and mechanism-simplicity of PID to meet the need of real-time control. This PIDNN was used to the steward tracking control system.