抄録
We present a framework for handling disturbances in a humanoid robot during walking motion. Based on the simplified dynamics of a mass concentrated model, we adopt a two level control strategy based on the actual CoM state space. The switching criteria is determined by the extreme values of the computed ZMP. If the extreme values are within the safety limit, the disturbance can be locally compensated by modifying the trajectory to the goal state. In the other case, the final state is modified. The proposed adaptive control is based on our previously presented walking motion generation scheme based on minimal energy control.
