The Control Research Of 6PUS-UPU Redundantly Actuated Parallel Robot Based On Neural Network

With the social demands changed rapidly, the requirements of the parallel robot control system become more and more demanding. This paper achieved a more accurate force position hybrid control system based on 6PUS-UPU redundant parallel robot. This paper solved the internal forces of the mechanical system and achieve a closed-loop control system for parallel robot, the paper research the intelligent control algorithm and conducted the simulation and experiment of the system, and achieved the online-adjust control of 6PUS-UPU redundant parallel robot, and verified the effectiveness of the control system.Research from the perspective of robot control requirements, this paper established the dynamic model of driving slides, driving branches, middle branch of Redundant mechanism and the whole parallel robot by Kane method, and then solved and optimized the driving force and internal force of the ends of driving branches. And simulated the dynamic model of the parallel robot through ADAMS, verified that the redundant driving dynamic model is correct.In view of the control method, the paper researched the BP neural network, and combined BP algorithm with conventional PID method based on the advantage of the BP neural network, and the control system can adjust those three parameters of PID online. The paper built mathematical models for non-redundant branch servo system and the transmission mechanism, and simulated the dynamic of non-redundant branch through conventional PID control method and PID control method based on BP neural network, verified the feasibility of applying PID control algorithm based on BP neural network to parallel robot control.In view of the control strategy, this paper applied force position hybrid control strategy to redundant mechanism, and used position control in non-redundant branches, and applied internal force control which added force feed-forward to redundant branch. The paper built the model of non-redundant branch servo system, and established the ADAMS-MATLAB co-simulation model of the redundant mechanism, conducted the co-simulation of conventional PID control method and PID control method based on BP neural network for parallel robot in condition that with and without interference, verified the PID control system based on BP neural network is more accurate and stable compared to conventional PID control.This paper conducted experiment research for 6PUS-UPU parallel robot, and verified that the parallel robot is more stable and have a better accuracy in position following when the PID control method based on BP neural network applied to it.