Zero-Shot Sim-to-Real Robot Learning: A Dexterous Manipulation Study on Reactive Catching
This work addresses the problem of sim-to-real transfer for highly sensitive dexterous manipulation tasks, offering a method that reduces the need for real-world fine-tuning.
The paper tackles the challenge of zero-shot sim-to-real transfer for dexterous manipulation, specifically reactive catching with a flat plate. The proposed Domain-Randomized Instance Set (DRIS) method achieves reliable transfer without real-world fine-tuning, demonstrating robust policies that handle uncertainties.
Dexterous manipulation is physics-intensive and highly sensitive to modeling errors and perception noise, making sim-to-real transfer prohibitively challenging. Domain randomization (DR) is commonly used to improve the robustness of learned policies for such tasks, but conventional DR randomizes one instance per episode, offering very limited exposure to the variability of real-world dynamics. To this end, we propose Domain-Randomized Instance Set (DRIS), which represents and propagates a set of randomized instances simultaneously, providing richer approximation of uncertain dynamics and enabling policies to learn actions that account for multiple possible outcomes. Supported by theoretical analysis, we show that DRIS yields more robust policies and alleviates the need for real-world fine-tuning, even with a modest number of instances (e.g., 10). We demonstrate this on a challenging reactive catching task. Unlike traditional catching setups that use end-effectors designed to mechanically stabilize the object (e.g., curved or enclosing surfaces), our system uses a flat plate that offers no passive stabilization, making the task highly sensitive to noise and requiring rapid reactive motions. The learned policies exhibit strong robustness to uncertainties and achieve reliable zero-shot sim-to-real transfer.