Cutting Sequence Diffuser: Sim-to-Real Transferable Planning for Object Shaping by Grinding

2024-12-19

IEEE Robotics and Automation Letters

AI Summary

In robotic grinding, accurate shape evolution modeling is hindered by strong coupling among material resistance, material removal volume, and tool pose, while real-world data collection is costly and irreversible. To address this, we propose a Sim-to-Real long-horizon grinding planning method relying solely on inexpensive synthetic simulation data. Our key contributions are: (1) a constrained, smooth action space enforcing small incremental material removal, explicitly mitigating dynamics discrepancies between simulation and reality; and (2) the first diffusion-model-based framework for generating physically feasible, long-sequence grinding trajectories, enabling efficient global planning. Experiments demonstrate high-precision grinding across diverse materials and complex target geometries, confirming strong cross-material and cross-shape generalization capability, as well as real-time deployability.

Technology Category

• Intelligent Robots: Motion and Path Planning
• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Computer Vision: Diffusion Models for Vision

Application Category

• Web Mining and Content Analysis: Web data generation and simulation
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Automating object shaping by grinding with a robot is a crucial industrial process that involves removing material with a rotating grinding belt. This process generates removal resistance depending on such process conditions as material type, removal volume, and robot grinding posture, all of which complicate the analytical modeling of shape transitions. Additionally, a data-driven approach based on real-world data is challenging due to high data collection costs and the irreversible nature of the process. This letter proposes a Cutting Sequence Diffuser (CSD) for object shaping by grinding. The CSD, which only requires simple simulation data for model learning, offers an efficient way to plan long-horizon action sequences transferable to the real world. Our method designs a smooth action space with constrained small removal volumes to suppress the complexity of the shape transitions caused by removal resistance, thus reducing the reality gap in simulations. Moreover, by using a diffusion model to generate long-horizon action sequences, our approach reduces the planning time and allows for grinding the target shape while adhering to the constraints of a small removal volume per step. Through evaluations in both simulation and real robot experiments, we confirmed that our CSD was effective for grinding to different materials and various target shapes in a short time.

Problem

• Automating robot grinding for object shaping
• Modeling complex shape transitions from removal resistance
• Reducing reality gap with sim-to-real transfer

Innovation

• Sim-to-real transferable planning using diffusion model
• Constrained small removal volumes for smooth action space
• Efficient long-horizon sequence generation from simulation data