D.J. Braun and M. Goldfarb, A controller for dynamic walking in bipedal robots, IEEE/RSJ International Conference on Intelligent Robots and Systems, St. Louis, USA, pp. 2916–2921, October 2009.
This paper presents a closed-loop control method for bipedal robots that achieves natural-looking and energy-efficient walking without prescribing joint trajectories or kinematic constraints. The approach uses state-dependent torques, some referenced to the inertial frame, which decouple angular dynamics and simplify controller tuning. Simulations on an anthropomorphic biped show rapid convergence to stable walking across different conditions, with walking speed adjustable through intuitive control parameters. The method also demonstrates low mechanical cost of transport, confirming its efficiency.
Why it matters: Traditional trajectory-tracking methods for biped robots often lead to rigid and inefficient motion. This work shows that encouraging patterned movement through state-dependent torques can yield more natural and efficient walking, influencing the development of energy-efficient humanoid robots.