The Research Of FNN,WNN And Its Application To Flexible Redundant Robots Control

Robots in the new era are desired to be light, high-speed, precise and high- flexible. Flexible redundant manipulators which possess characters of both flexible and redundant manipulators, have extensive applied prospects in aeronautical fields. However, these manipulators are difficult to precisely control because of vibration caused by the structural flexibility. So, fixing position of target and reducing the residual vibration quickly are the two tasks for the control system of flexible redundant manipulators. With the idea of treating the rigidity and flexibility respectively, this paper studies how to design a distributed control system for the flexible redundant manipulators, which fix position of target under the control strategy for the rigid manipulators, and reduce the vibration under the active control for the flexible ones. Based on the research into the principium, structure and algorithms of the fuzzy neural networks (FNN) and wavelet neural networks (WNN), the Author conceive to finish the tasks of fixing position of target and reducing the residual vibration by the FNN controller and WNN controller respectively. During the period of control, FNNC and WN7NC work at the different time, which means that FN7NC works in the process of fixing position of target, and WNNC will switch to work to reduce the residual vibration at the time when the process is over. Lots of simulating experiments indicate that, the control system for flexible and redundant manipulators with FN7NC and WNNC is feasible, the precision of fixing position of target and reducing the residual vibration is contentment.