Fabrica: Dual-Arm Assembly of General Multi-Part Objects via Integrated Planning and Learning

2025-06-05

AI Summary

Existing approaches for long-horizon, contact-intensive assembly of multi-component objects suffer from limited geometric generality and insufficient autonomy. Method: We propose the first end-to-end bimanual autonomous assembly framework that requires neither domain-specific knowledge nor human demonstrations. Our approach integrates long-horizon hierarchical symbolic planning—reasoning over assembly sequences, precedence constraints, and grasp configurations—with contact-aware lightweight reinforcement learning control. Key innovations include an equivariance-guided residual action policy, automatic gripper generation, and parallelized motion planning. Contribution/Results: The framework enables zero-shot cross-object generalization and achieves 100% complete assembly success on a real-world benchmark spanning industrial and household objects. It attains an 80% success rate in contact-intensive manipulation tasks and has been successfully deployed on a physical bimanual robot platform.

Technology Category

• Intelligent Robots: Manipulation
• Humans and AI: Human-Aware Planning and Behavior Prediction
• Planning, Routing, and Scheduling: Learning for Planning and Scheduling

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Multi-part assembly poses significant challenges for robots to execute long-horizon, contact-rich manipulation with generalization across complex geometries. We present Fabrica, a dual-arm robotic system capable of end-to-end planning and control for autonomous assembly of general multi-part objects. For planning over long horizons, we develop hierarchies of precedence, sequence, grasp, and motion planning with automated fixture generation, enabling general multi-step assembly on any dual-arm robots. The planner is made efficient through a parallelizable design and is optimized for downstream control stability. For contact-rich assembly steps, we propose a lightweight reinforcement learning framework that trains generalist policies across object geometries, assembly directions, and grasp poses, guided by equivariance and residual actions obtained from the plan. These policies transfer zero-shot to the real world and achieve 80% successful steps. For systematic evaluation, we propose a benchmark suite of multi-part assemblies resembling industrial and daily objects across diverse categories and geometries. By integrating efficient global planning and robust local control, we showcase the first system to achieve complete and generalizable real-world multi-part assembly without domain knowledge or human demonstrations. Project website: http://fabrica.csail.mit.edu/

Problem

• Dual-arm robotic system for autonomous multi-part assembly
• Long-horizon planning with hierarchies for general assembly
• Lightweight reinforcement learning for contact-rich assembly steps

Innovation

• Hierarchical planning with automated fixture generation
• Lightweight reinforcement learning for contact-rich assembly
• Integrated global planning and local control