Dynamics as Prompts: In-Context Learning for Sim-to-Real System Identifications
This addresses the problem of precise robot control in dynamic real-world scenarios, offering an incremental improvement over traditional domain randomization methods.
The paper tackles the challenge of sim-to-real transfer in robotics by proposing an in-context learning approach that dynamically adjusts simulation parameters online using past interaction histories, achieving 80% and 42% improvements in parameter estimation over baselines in sim-to-sim tasks and at least 70% success rate in sim-to-real object scooping.
Sim-to-real transfer remains a significant challenge in robotics due to the discrepancies between simulated and real-world dynamics. Traditional methods like Domain Randomization often fail to capture fine-grained dynamics, limiting their effectiveness for precise control tasks. In this work, we propose a novel approach that dynamically adjusts simulation environment parameters online using in-context learning. By leveraging past interaction histories as context, our method adapts the simulation environment dynamics to real-world dynamics without requiring gradient updates, resulting in faster and more accurate alignment between simulated and real-world performance. We validate our approach across two tasks: object scooping and table air hockey. In the sim-to-sim evaluations, our method significantly outperforms the baselines on environment parameter estimation by 80% and 42% in the object scooping and table air hockey setups, respectively. Furthermore, our method achieves at least 70% success rate in sim-to-real transfer on object scooping across three different objects. By incorporating historical interaction data, our approach delivers efficient and smooth system identification, advancing the deployment of robots in dynamic real-world scenarios. Demos are available on our project page: https://sim2real-capture.github.io/