| Nowdays, force/position control for robot manipulators is one of the challenging subjects of developing robot technology. In the recent decades, many efforts have been devoted to the control of robot manipulators, especially, problems of free motion and constrained motion control for robot manipulators have attached many researchers. It is know that there exist complex nonlinear, strong coupling and lots of uncertains in robotic system. When the manipulator end-effector contacts with the environment, the different environment stiffness effect greatly the system performance. Hence, good adaptive and strong robustness is required in robotic system. In this dissertation, several problems of force and position control for robot manipulators have been studied in the background of robot practical tasks, the main results are as follows:The dissertation gives a brief description about the movement control from the direction of work-space. This dissertation presents tow controllers about robot adaptive control based on NN for the uncertains of robot and the environment. Adaptive control compensate for uncertains of parameters, robust control compensate for uncertains of non-parameters in these controllers, RBF is used in the one, GA combined with the BP algorithms of neural networks is presented in the other. The latter algorithm can search the best weights of neural networks in whole searching space. Robust adaptive force/position hybrid control based on neural networks is presented, a comprehensive simulation study has been carried out using a two-degree of freedom contacted with the environment. To the character of tasks, two impendence structure are suggested, making up for the shortcoming of hybrid control, NN is used to leaning the uncertains. The two schemes have good robust and high value in practice . We focus on the six degree parallel robot model made by Yanshan University. On the basis of former work, the result of the simulation experiment is satisfying.
|
PDF Full Text Request