Recovery control by using visually estimated foot sole floating angle for biped walking robot

The paper presents a recovery control method based on visually estimated foot sole floating angle for biped walking robots. By using visual scene recognition through image processing, the floating angle of foot sole can be indirectly estimated, and that means the critical state of falling-down motion is able to be detected without force sensation. According to estimated foot sole floating angle, the moment generator is designed in order to recover from floating state of foot sole in this paper. The proposed method will contribute to the improvement of recovery performance against disturbances while standing conditions. The validity of the proposed method was evaluated through several experiments.