Reflex and gait adaptation for biped walking robots in phase of changing payloads

Biped walking robots present one of the most suitable mechanical configurations to work in the human environment. They have significant advantages in obstacle avoidance when compared with their wheeled counterparts. However, their control is a challenging problem due to their inherent instability and non-linearity. Recently, intensive research is being conducted to employ biped walking robots in heavy and dangerous tasks of human operators. Carrying heavy loads is one of the natural tasks a biped walking robot of future might be expected to accomplish. Varying payload, however, can change the robot dynamics and cause instability. This paper proposes a walking pattern adaptation technique to compensate for sudden additions of loads of unknown masses. The proposed scheme is tested on a full dynamics simulation test bed, with simulations of a 12 dof walking robot.