Falling Prevention System from External Disturbances for Humanoid Robots

Humanoid robot requires a robust prevention system against external disturbances to protect itself from falling to the ground and to perform its tasks completely. In this paper, a Falling Prevention System for humanoid robot is proposed to avoid falling from the disturbances, and helps humanoid robot recover its balance from external force by taking a step. The algorithm for the Falling Prevention System consists of two processes. First, humanoid robot can perceive whether it is falling or not by using an IMU sensor, and if falling, the center of mass (CoM) and swinging leg trajectories are calculated for the robot to take a step. The CoM and swinging leg trajectories are also used to acquire all joint angles of lower body by inverse kinematics. Furthermore, designed foot trajectory helps humanoid robot minimize its yawing moment. Next, mass-spring-damper system for the robot's legs is modeled to reduce large impact force from the ground. The effectiveness of the proposed method is demonstrated through computer simulations for a humanoid robot.