Lazy control strategy for active passive walking

Passive walking on a slope is known as a most efficient walking style because it needs no external power, and the motion occurs naturally according to its mechanical properties. We introduce the idea of "virtual incline gravity" and show that a walker on level ground can walk from an adequate initial state and an equivalent torque which corresponds to the virtual gravity effect. Passive walking has a stable limit cycle, but the walker can\´t continue walking under the existence of parameter uncertainty or disturbances because the stable region of the trajectory is narrow. To enhance robustness, we propose a \´lazy control\´ strategy, and apply this method to passive-like walking control. The strategy is based on the concept that an input is added to the system only when the state gets outside the stable region. By using this approach, the walker can walk by itself as long as its own stability is valid without any external input, and is helped to recover stable motion only when the stability is likely to vanish. The effectiveness of this strategy is demonstrated by simulation