3 accepted papers
We introduce SoftMimic , a framework for learning compliant whole-body control policies for humanoid robots from example motions. Imitating human motions with reinforcement learning allows humanoids to quickly learn new skills, but existing methods incentivize stiff control that aggressively correct…
Training controllers via reinforcement learning (RL) in simulation has emerged as a powerful approach for synthesizing robust and agile robotic behaviors evaluated in reality. We push the envelope of the simulation training paradigm by exposing problems encountered when learning agile behaviors only…
As quadruped controllers approach greater maturity for locomotion on level ground, a next challenge relates to enabling these systems to carefully choose contacts in cluttered or discontinuous terrain. In pursuit of this goal, this paper proposes an approach to motion generation for dynamic hopping