Reflexive Control of Slipping in an Anthropomorphic Bipedal Locomotion Robot

Abstract

To be successful a bipedal locomotion system will be required to robustly traverse terrain with unknown and varying surface conditions. Even in made-for-human environments a broad range of surface conditions may be encountered and slipping is a very real and serious concern. Since a priori detection and avoidance is not practical, a robust system must detect and respond to the unexpected event after it has occurred and before control is lost. This must be accomplished in an extremely limited time frame. This paper proposes the use of human-like reflexes to enable robust recovery from slipping. The proposed algorithm seamlessly integrates reflexive responses into the nominal control system. Effectiveness of the algorithm is demonstrated via simulation results.