Generalizable Friction Coefficient Estimation via Material Embedding and Proxy Interaction Modeling

TL;DR

This paper presents a proxy-based learning framework for estimating friction coefficients between arbitrary materials efficiently, reducing the need for exhaustive pairwise testing and enabling accurate, interpretable predictions with uncertainty estimates.

Contribution

The authors introduce a novel material embedding and proxy interaction model that accurately predicts friction coefficients with fewer tests and handles noisy or missing data.

Findings

Achieves high predictive accuracy on simulated and real datasets.

Reduces experimental testing by using proxy materials.

Provides calibrated uncertainty estimates for friction predictions.

Abstract

Accurately estimating friction coefficients between arbitrary material pairs is critical for robotics, digital fabrication, and physics-based simulation, but exhaustive pairwise testing scales quadratically with the number of materials. We introduce a proxy-based modeling framework that approximates any pairwise friction $f(A,B)$ from a small, fixed set of proxy materials $C=[c_1,\dots,c_k]$ by learning a per-material embedding $z_A = g(f(A,c1),\dots,f(A,ck))$ and a fusion function $p$ such that $f(A,B)\approx p\big(z_A,z_B\big)$. We present deterministic and probabilistic realizations of $g$ and $p$, procedures for selecting diverse proxy sets, and mechanisms for handling missing or noisy proxy measurements. The learned embeddings are compact, interpretable, and enable calibrated uncertainty estimates for downstream decision making. On simulated and measured friction datasets, our approach achieves high predictive accuracy, robust performance with partial observations, and substantial experimental savings by significantly reducing pairwise testing.