Trademark 3840-480

A control framework based on center-of-mass momentum distribution: improving the walking stability of small humanoid robots

Most small humanoid robots are unable to detect ground conditions accurately and can barely walk on uneven terrain. To solve this problem, Zhicheng He's research team from China has proposed a walking control framework based on center-of-mass momentum distribution to help improve the motion stability of small humanoid robots. The control framework consists of three controllers: a momentum decrement controller, a pose controller, and a guide controller, and uses a MEMS-based IMU (inertial measurement unit) and a joint position encoder to calculate the center-of-mass momentum.

To evaluate the effectiveness of the control framework, the team conducted three types of experiments to test standing balance, discrete terrain walking and uneven terrain walking.
 In the standing balance experiment, the humanoid robot could recover from the impact of 40ON/0.1s; in the discrete terrain walking experiment, the walking success rate of the humanoid robot improved from 40% to about 90%; in the uneven terrain walking experiment, the humanoid robot walked successfully on the rugged ground. The experimental results demonstrate the effectiveness of momentum distribution control, and the walking control framework based on center-of-mass momentum distribution can effectively improve the anti-disturbance control ability and walking stability of the small humanoid robot.