Mates2Motion: Learning How Mechanical CAD Assemblies Work

TL;DR

This paper introduces Mates2Motion, a deep learning approach that infers degrees of freedom in mechanical CAD assemblies, improving understanding of part interactions and motion prediction.

Contribution

It presents a novel deep learning method trained on a large dataset to infer and refine mates and axes of motion in CAD assemblies, enhancing motion analysis accuracy.

Findings

Effective inference of degrees of freedom in assemblies

Improved methods for redefining mates for better motion reflection

Creation of a reliable, motion-annotated test set

Abstract

We describe our work on inferring the degrees of freedom between mated parts in mechanical assemblies using deep learning on CAD representations. We train our model using a large dataset of real-world mechanical assemblies consisting of CAD parts and mates joining them together. We present methods for re-defining these mates to make them better reflect the motion of the assembly, as well as narrowing down the possible axes of motion. We also conduct a user study to create a motion-annotated test set with more reliable labels.