Do Rigid-Body Simulators Dream of Soft Robots? Learning Contact-Rich Manipulation for Tendon-Driven Continuum Robots

2026-06-21

AI Summary

This work addresses the lack of high-fidelity simulation infrastructure for soft continuum robots capable of contact-rich manipulation, which has hindered the development of learning-based whole-body control. The authors propose a discretized modeling approach grounded in continuum mechanics that natively embeds tendon-driven continuum robots (TDCRs) into MuJoCo, unifying tendon actuation, self-contact, and dynamics within a single physics simulation framework. Leveraging teleoperation data for policy training, the method enables zero-shot transfer to physical hardware. The simulation exhibits strong agreement with Cosserat rod theory in both static and dynamic scenarios and successfully executes two contact-intensive tasks on a real three-segment TDCR. This represents the first demonstration of zero-shot sim-to-real transfer for continuum robots, bridging a critical gap between high-fidelity simulation and learning-based control in soft robotics.

Abstract

Learning contact-rich, whole-body manipulation for soft continuum robots is held back by the lack of simulation infrastructure that has accelerated rigid-robot manipulation. Existing soft robot simulators are physically grounded but lack the contact handling, actuation support, or learning integration needed for contact-rich manipulation; rigid-body approximations offer these capabilities but sacrifice physical grounding. We bridge this gap for tendon-driven continuum robots (TDCRs) by deriving a continuum-mechanics-informed discretization that places the soft robot natively inside MuJoCo, unifying tendon forces, body contact, and dynamics in a single physics pipeline. We validate the simulator against a Cosserat rod reference (static and dynamic) and real TDCR hardware. We then train state-based imitation learning policies via teleoperation in simulation and deploy them zero-shot to a physical 3-segment TDCR on a 7-DoF Franka arm across two contact-rich manipulation tasks. To our knowledge, this is the first demonstration of sim-to-real transfer for contact-rich manipulation with continuum robots.

Problem

• soft robots
• contact-rich manipulation
• simulation infrastructure
• tendon-driven continuum robots
• sim-to-real transfer

Innovation

• continuum robots
• tendon-driven
• MuJoCo simulation
• contact-rich manipulation
• sim-to-real transfer