Abstract
This study deals with harnessing for a quadruped robot toward seeing-eye dog robot. The control system is composed of two blocks: First one is spontaneous locomotion control system that is a oscillator network with phase-resetting. And the other is recognition system for human-machine interface that is a neural network to distinguish commanded motion pattern or harnessing pattern from human. Efficiency of the proposed control system is verified through numerical simulations and hardware experiments.
