CLASH: Collision Learning via Augmented Sim-to-real Hybridization to Bridge the Reality Gap

2026-02-20

AI Summary

This work addresses the challenge of policy failure in simulation-to-reality (sim-to-real) transfer caused by inaccurate collision dynamics modeling. To this end, the authors propose the CLASH framework, which first distills fundamental physical priors from MuJoCo and then fine-tunes a lightweight surrogate collision model using only ten real-world interaction samples to construct a high-fidelity hybrid simulator. This approach substantially improves simulation accuracy while reducing collision computation overhead by nearly 50%. Evaluated on real-world continuous object-pushing tasks, reinforcement learning policies trained with CLASH achieve double the success rate compared to baselines, and model predictive control performance is also significantly enhanced, demonstrating data-efficient and high-fidelity sim-to-real transfer.

Technology Category

• Machine Learning: Imitation Learning & Inverse Reinforcement Learning
• Search and Optimization: Sampling/Simulation-based Search
• Planning, Routing, and Scheduling: Replanning and Plan Repair

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
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

The sim-to-real gap, particularly in the inaccurate modeling of contact-rich dynamics like collisions, remains a primary obstacle to deploying robot policies trained in simulation. Conventional physics engines often trade accuracy for computational speed, leading to discrepancies that prevent direct policy transfer. To address this, we introduce Collision Learning via Augmented Sim-to-real Hybridization (CLASH), a data-efficient framework that creates a high-fidelity hybrid simulator by learning a surrogate collision model from a minimal set of real-world data. In CLASH, a base model is first distilled from an imperfect simulator (MuJoCo) to capture general physical priors; this model is then fine-tuned with a remarkably small number of real-world interactions (as few as 10 samples) to correct for the simulator's inherent inaccuracies. The resulting hybrid simulator not only achieves higher predictive accuracy but also reduces collision computation time by nearly 50\%. We demonstrate that policies obtained with our hybrid simulator transfer more robustly to the real world, doubling the success rate in sequential pushing tasks with reinforecement learning and significantly increase the task performance with model-based control.

Problem

• sim-to-real gap
• collision dynamics
• contact-rich interactions
• policy transfer
• physics simulation

Innovation

• sim-to-real transfer
• collision modeling
• hybrid simulation
• data-efficient learning
• physics distillation