AutoURDF: Unsupervised Robot Modeling from Point Cloud Frames Using Cluster Registration

2024-12-07

arXiv.org

AI Summary

This work addresses the reliance on manual annotation and prior knowledge in modeling unknown robots. We propose the first unsupervised, end-to-end, point-cloud-driven framework for automatic URDF generation. Given multi-frame unannotated point clouds, our method jointly performs part segmentation, hierarchical topology inference, and joint parameter estimation via motion-consistency-driven clustering, registration, and 6-DoF rigid-body tracking—ultimately producing executable, Gazebo/PyBullet-compatible URDF models. Key contributions are: (1) full automation without human intervention or robot-specific priors; and (2) the first motion-guided, unsupervised mechanism for topology inference. Evaluated on both synthetic and real-world scanned data, our approach achieves a 12.3% improvement in registration accuracy and an 18.7% gain in topology identification accuracy, significantly reducing modeling effort and enhancing feasibility of simulation deployment.

Technology Category

• Intelligent Robots: State Estimation
• Reasoning under Uncertainty: Relational Probabilistic Models
• Computer Vision: Motion & Tracking

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Robot description models are essential for simulation and control, yet their creation often requires significant manual effort. To streamline this modeling process, we introduce AutoURDF, an unsupervised approach for constructing description files for unseen robots from point cloud frames. Our method leverages a cluster-based point cloud registration model that tracks the 6-DoF transformations of point clusters. Through analyzing cluster movements, we hierarchically address the following challenges: (1) moving part segmentation, (2) body topology inference, and (3) joint parameter estimation. The complete pipeline produces robot description files that are fully compatible with existing simulators. We validate our method across a variety of robots, using both synthetic and real-world scan data. Results indicate that our approach outperforms previous methods in registration and body topology estimation accuracy, offering a scalable solution for automated robot modeling.

Problem

• Automates robot modeling from point clouds
• Identifies moving parts and joint parameters
• Generates simulator-compatible robot description files

Innovation

• Unsupervised robot modeling from point clouds
• Cluster-based 6-DoF transformation tracking
• Automated URDF generation for simulators