Search papers, labs, and topics across Lattice.
This paper introduces a contact-implicit, direct simulation-based multiple shooting (DSMS) framework that enhances motion retargeting by transforming kinematically feasible references into dynamically feasible whole-body trajectories. By integrating a differentiable simulator within a nonlinear program, the method effectively manages contact dynamics, friction, and actuation limits without explicit contact constraints, leading to improved reinforcement learning (RL) policy training. The results show that DSMS significantly accelerates motion-imitation RL training and achieves high success rates with low tracking errors, including successful zero-shot sim-to-real transfer on a physical robot.
Transforming kinematically feasible motion references into dynamically accurate trajectories could revolutionize how robots learn complex contact-rich behaviors.
Motion retargeting approaches often prioritize kinematic similarity over whole-body dynamics, contact consistency, and actuation limits, yielding references that are difficult for reinforcement learning (RL) policies to reproduce, particularly for contact-rich behaviors. We present a contact-implicit, direct simulation-based multiple shooting (DSMS) framework that transforms kinematically feasible references into dynamically feasible whole-body trajectories. By embedding a differentiable simulator within a nonlinear program, DSMS resolves contact, friction, impacts, self-collision, and joint limits internally while enforcing tracking, actuation, and task constraints without prescribing a contact schedule or introducing explicit contact constraints. Compared with existing retargeting methods, DSMS accelerates motion-imitation RL training and yields policies with high success rates and low tracking error. We further demonstrate zero-shot sim-to-real transfer on the Unitree G1 through command-conditioned contact-rich crawling and a highly dynamic 180-degree jump-turn.
