Standing Balance Control for Position Control-Based Humanoid Robot

This work presents a standing balance control method against external disturbance for the position control-based humanoid robot. The method consists of a simple but effective posture estimator, a state machine based on capture point and a virtual model exhibiting the variable compliance. The posture estimator appraises the influence of external disturbances. The state machine switches the strategies for humanoid push recovery. The virtual model applied on position-based robot enables the robot to have both obedience and resistance to the external force in the stable state. The proposed method was tested on a position control-based humanoid robot “KONG” with 14 degrees of freedom. Experiments and related data show the humanoid robot “KONG” has perfect balance performance for various external force disturbances.