Exploring the dynamic walking range of the biped robot "Run Bot" with an active upper-body component

Abstract

In this paper, we explore the dynamic walking capability of the planar biped robot RunBot with a now added active upper-body component. Originally, the robot was designed and built to perform fast walking especially on a flat floor. Its locomotion is driven by so-called neural reflexive control. This controller does not employ any kind of position or trajectory-tracking control algorithm. Instead, it enables RunBot to exploit its own natural dynamics during critical stages of its gait cycles. The actual gait pattern is determined by the set of neural control parameters, like gain and activation thresholds. Thus, different gait patterns can be induced by changing these parameters. These walking patterns, cooperating with an added active-upper body component, allow RunBot to walk on different terrains, e.g. flat floor, up and down slopes between 0 and 7.5 degrees. The transition phase of each gait was experimentally tuned. As a result, RunBot can continuously walk on the three different terrains. During walking experiments, gait switching was manually triggered while the active body was controlled to lean either forward or backward according to the slope.