Gait Transition Method for Quadruped Robot Based on CPG Network and Impedance Control

This paper concerns the problem of Central Pattern Generator (CPG) applying on large-scale quadruped robot control. We found that trajectories generated by the CPG model show sharp inflections points between steps, and this kind of inflections will have more negative effect on larger robot locomotion because of the scale-effect. We proposed a control architecture in which a Hopf oscillator is high-level controller while low-level is impedance controller. In this architecture, Hopf oscillator generates motion rhythm and motion patterns which can be executed smoothly by impedance controller. Finally, we apply the control method on a simulated quadruped robot to verify the effectiveness. The results show that the robot can achieve smooth and stable locomotion.