Abstract
Quadrupeds exhibit adaptive locomotion under real-world environment. Such locomotion is generated via the coordination between legs, i. e., interlimb coordination and is controlled by a neural network called central pattern generator (CPG). However, underlying mechanism remains elusive. To establish the design principle of highly adaptive quadruped locomotion, we propose a "TEGOTAE"-based CPG control that exploits vertical and horizontal ground reaction forces (GRFs) from environment. In this paper, we develop a quadruped robot with GRF sensing mechanism and verify the validity of proposed control scheme.
