| The6-DOF manipulator system is a strongly coupled nonlinear multiple-in-put-multiple-output (MIMO) plant, which is widely applied in the industrial field of painting, welding and automatic handling. It is important to control the location of the end effector. So a decoupled method based on inverse system is proposed to design the controller and realize the rapid and precise control for the end effector in this dissertation.The inverse system of the6-DOF manipulator is constructed by mathematical reasoning and support vector machine-based (SVM) recognition. What’s more, genetic algorithm (GA) can be used to optimize support vector machine parameters. Simulation results show that the6-DOF manipulator is equivalent to six single-input-single-output (SISO) integral chains. Although the later decoupling effect is not as desirable as the former, the fussy derivation process can be avoided.The inverse system is connected serially to the plant, which leads to a pseudo linear system. In the fact, it is hard to ensure accuracy of the pseudo linear system. So a feedback controller which is based on internal model principle is designed to realize the control for position of each joint. There are several filter parameters in the controller. Each joint will be controlled just through adjusting the filter parameters. Its stability is analyzed and discussed. Experimental results illustrate the effectiveness when the IMC (internal model controller) which is based on SVM inverse system method is used to control the6-DOF manipulator system. The location of each joint can reach the given value. Compared to PID controller, it has smaller error and better robustness, what’s more, the former has less parameters need to adjust and its joints can be controlled separately. In practical engineering application, engineers need spend too much time on parameter adjustment to ensure the control accuracy. Therrfore, if the former controller is selected, the adjustment time will be reduced significantly. |
PDF Full Text Request