Dynamic Simulation and Analysis of a Passively Self-Stabilizing Hexapedal Running Robot

Abstract

The Sprawl family of robots have demonstrated that fast and robust locomotion is possible over uneven terrain with only feed-forward control. In the absence of feed-back, the speed, efficiency, and stability of the system are largely functions of how well the physical system is tuned. This work describes the development of a dynamic model used to investigate the effects of design changes on system stability and performance. Simulation results show the surprisingly large region of configurations that result in stable locomotion, and how the shape of this region changes as a function of the passive properties of the hips. Understanding of this is critical since peak performance, as measured by running speed, tends to occur on the boundary of stable configurations. Simulation results were used to quantitatively redesign the legs of a version of the robot configured to run outdoors. Implementation of the design changes resulted in doubling the robot’s speed.