Surrogate compliance modeling enables reinforcement learned locomotion gaits for soft robots

2025-12-07

AI Summary

Addressing the significant simulation-to-reality gap, high computational cost, and lack of high-fidelity soft-body simulators in soft robotics, this paper proposes a surrogate compliance modeling approach: it introduces indirect variables representing soft deformation into rigid-body dynamics simulation, and integrates reinforcement learning with domain randomization to enable efficient learning and cross-domain transfer of closed-loop gait policies. This method achieves, for the first time, high-fidelity soft-body motion control within a purely rigid-body simulator, overcoming the longstanding trade-off between accuracy and efficiency in soft robotics simulation. The learned policies are successfully deployed on a biomimetic amphibious turtle-inspired robot, enabling stable, multimodal locomotion across diverse natural terrains. On land, mobility is significantly enhanced, and the transport cost is reduced by an order of magnitude compared to open-loop baselines.

Technology Category

• Intelligent Robots: Embodied AI
• Machine Learning: Imitation Learning & Inverse Reinforcement Learning
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Adaptive morphogenetic robots adapt their morphology and control policies to meet changing tasks and environmental conditions. Many such systems leverage soft components, which enable shape morphing but also introduce simulation and control challenges. Soft-body simulators remain limited in accuracy and computational tractability, while rigid-body simulators cannot capture soft-material dynamics. Here, we present a surrogate compliance modeling approach: rather than explicitly modeling soft-body physics, we introduce indirect variables representing soft-material deformation within a rigid-body simulator. We validate this approach using our amphibious robotic turtle, a quadruped with soft morphing limbs designed for multi-environment locomotion. By capturing deformation effects as changes in effective limb length and limb center of mass, and by applying reinforcement learning with extensive randomization of these indirect variables, we achieve reliable policy learning entirely in a rigid-body simulation. The resulting gaits transfer directly to hardware, demonstrating high-fidelity sim-to-real performance on hard, flat substrates and robust, though lower-fidelity, transfer on rheologically complex terrains. The learned closed-loop gaits exhibit unprecedented terrestrial maneuverability and achieve an order-of-magnitude reduction in cost of transport compared to open-loop baselines. Field experiments with the robot further demonstrate stable, multi-gait locomotion across diverse natural terrains, including gravel, grass, and mud.

Problem

• Modeling soft robot compliance in rigid-body simulators for accurate simulation.
• Achieving reliable policy learning via reinforcement learning with surrogate variables.
• Enabling sim-to-real transfer of locomotion gaits across complex terrains.

Innovation

• Surrogate compliance modeling replaces soft-body physics simulation
• Indirect variables represent deformation in rigid-body simulator
• Reinforcement learning with variable randomization enables sim-to-real transfer