Active stabilization of a humanoid robot base on inertial measurement unit data

In this paper we describe the active stabilization of a humanoid robot. It is based on the data obtained from an Inertial Measurement Unit (IMU), which is placed in the robot's torso. IMU includes accelerometer and gyroscope. Relying on these data, we propose the algorithm that maintains stability of the robot, while the support platform tilts. Our experimental object was a small humanoid robot M-Bot, which was built by authors and is shortly described in this paper. To determine the orientation of the robot we developed a digital filter. To maintain the static stability we decided to implement a controller based on the position of the Center of Mass (COM) projection on the ground plane. To adjust COM position we use ankle and hip strategy. These strategies provide stabilization only in one plane. To improve side stability we propose the leg adjustments. This algorithm detects the possibility of exceeding outer boundary of the foot by COM and adjusts configuration of the leg to prevent against falling to one side. The developed algorithms were fully tested on a real construction of M-bot robot. Experimental results demonstrated that our implementation successfully stabilizes a humanoid robot.