Shooting for Contact: Contact-Implicit Multiple Shooting for Dynamic Motion Retargeting

2026-08-04

AI Summary

This work addresses the limitations of traditional motion retargeting methods, which often neglect full-body dynamics, contact consistency, and actuator constraints, thereby struggling to reproduce complex contact-rich behaviors. The authors propose an optimization framework that integrates implicit contact handling with Direct Simulated Multiple Shooting (DSMS), embedding a differentiable simulator to automatically manage contacts, friction, collisions, and joint limits—without requiring predefined contact schedules or explicit contact constraints. This approach represents the first integration of implicit contact dynamics with multiple shooting, enabling efficient generation of high-fidelity full-body dynamic trajectories and substantially improving the efficiency and generalization of reinforcement learning-based imitation training. Experiments demonstrate that the learned policies achieve high success rates and low tracking errors in simulation and enable zero-shot sim-to-real transfer on the Unitree G1 robot, successfully executing challenging tasks such as crawling and a 180-degree jump-turn.

Abstract

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.

Problem

• motion retargeting
• contact consistency
• dynamic feasibility
• reinforcement learning
• whole-body dynamics

Innovation

• contact-implicit optimization
• differentiable simulation
• multiple shooting
• dynamic motion retargeting
• sim-to-real transfer