Neural Network Type Grasping Control for Robot Hand
Mitsuo Wada and Hiroshi Endo
Industrial Products Research Institute, 1-1-4, Higashi Tsukuba-shi Ibaraki 305, Japan
A neural network-type control method to grasp an object and apply force uniformly using multi-joint hand is presented. The controller has lateral inhibition-type parallel processing and autonomously folding toward decreasing the energy of grasping state.