Impedance Optimization for Uncertain Contact Interactions Through Risk Sensitive Optimal Control

This letter addresses the problem of computing optimal impedance schedules for legged locomotion tasks involving complex contact interactions. We formulate regulation as a trade-off between disturbance rejection and measurement uncertainty. extend stochastic control algorithm known Risk Sensitive Control to take into account uncertainty propose formal way include such unknown locations. The approach can efficiently generate state trajectories along with local feedback gains, i.e. schedules. Extensive simulations demonstrate capabilities in generating meaningful stiffness damping modulation patterns before after interaction. For example, forces are reduced during early contacts, increases anticipate high impact event tracking is automatically traded-off increased stability. In particular, we show significant improvement performance jumping trotting simulated quadruped robot.