RAFL: Generalizable Sim-to-Real of Soft Robots with Residual Acceleration Field Learning

TL;DR

This paper introduces RAFL, a residual acceleration field learning framework that enhances soft robot simulation accuracy across different shapes and real-world applications, addressing the sim-to-real gap effectively.

Contribution

RAFL provides a transferable, element-level correction to simulators, improving generalization and accuracy in sim-to-real transfer for soft robots with minimal computational overhead.

Findings

RAFL achieves zero-shot improvements on unseen morphologies.

The method outperforms traditional system identification in transfer scenarios.

RAFL enables continual refinement during morphology optimization.

Abstract

Differentiable simulators enable gradient-based optimization of soft robots over material parameters, control, and morphology, but accurately modeling real systems remains challenging due to the sim-to-real gap. This issue becomes more pronounced when geometry is itself a design variable. System identification reduces discrepancies by fitting global material parameters to data; however, when constitutive models are misspecified or observations are sparse, identified parameters often absorb geometry-dependent effects rather than reflect intrinsic material behavior. More expressive constitutive models can improve accuracy but substantially increase computational cost, limiting practicality. We propose a residual acceleration field learning (RAFL) framework that augments a base simulator with a transferable, element-level corrective dynamics field. Operating on shared local features, the model is agnostic to global mesh topology and discretization. Trained end-to-end through a differentiable simulator using sparse marker observations, the learned residual generalizes across shapes. In both sim-to-sim and sim-to-real experiments, our method achieves consistent zero-shot improvements on unseen morphologies, while system identification frequently exhibits negative transfer. The framework also supports continual refinement, enabling simulation accuracy to accumulate during morphology optimization.