Rearrangement Planning for General Part Assembly

TL;DR

This paper introduces GPAT, a transformer-based model for rearrangement planning in general part assembly, capable of predicting precise part poses for unseen shapes in both CAD models and real-world scans.

Contribution

The paper presents GPAT, a novel transformer architecture that generalizes to unseen part and target shapes for assembly tasks.

Findings

GPAT accurately predicts part poses in diverse scenarios.

GPAT generalizes well to unseen shapes in CAD and real-world data.

Transformer-based approach improves assembly planning accuracy.

Abstract

Most successes in autonomous robotic assembly have been restricted to single target or category. We propose to investigate general part assembly, the task of creating novel target assemblies with unseen part shapes. As a fundamental step to a general part assembly system, we tackle the task of determining the precise poses of the parts in the target assembly, which we we term ``rearrangement planning''. We present General Part Assembly Transformer (GPAT), a transformer-based model architecture that accurately predicts part poses by inferring how each part shape corresponds to the target shape. Our experiments on both 3D CAD models and real-world scans demonstrate GPAT's generalization abilities to novel and diverse target and part shapes.