Position and force hybrid control of robotic manipulator by neural network. (1st Report, Application of neural servo controller to stabbing control).

Abstract

In this paper, a new concept of a "neural servo controller" is presented to show the applicability of the neural network to the position and force control of robotic manipulators. The proposed neural servo controller is based on the self-organization capability of the neural network, which here consists of two hidden layers, and input/output layers. The controller can adjust the neural network ou-put to the robot in the forward manner to cooperate with the feedback loop, depending on different characteristics of handling objects. In particular, the neural network can recognize the force used to adjust the hybrid control ratio between the position and the force control modes. The proposed method can adapt the network to stabbing control as one of applications of the position and force control. Simulations as well as experiments are carried out for the case of one-dimensional robotic manipulators. The results show the applicability and adaptability of the proposed neural servo controller to the position/force control of manipulators.