FoldPath: End-to-End Object-Centric Motion Generation via Modulated Implicit Paths

2025-11-03

AI Summary

Existing object-centric motion generation (OCMG) methods rely on heuristic rules or learned models but require sensitive post-processing to produce executable trajectories, compromising continuity, smoothness, and robustness under complex 3D geometries. This work introduces the first end-to-end neural field framework for OCMG, modeling robot motion as a continuous implicit path function that directly outputs differentiable, seam-free trajectories—eliminating discrete point sampling and post-hoc refinement. Our approach integrates modulated implicit neural fields, deep trajectory representation learning, and continuous path optimization, augmented by a long-horizon path evaluation metric. Trained on only 70 expert demonstrations, it significantly outperforms state-of-the-art methods in simulation: generated trajectories exhibit superior smoothness, temporal coherence, and generalization across unseen objects and scenes. This advances OCMG toward industrial-grade practicality.

Technology Category

• Intelligent Robots: Motion and Path Planning
• Computer Vision: Motion & Tracking
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Object-Centric Motion Generation (OCMG) is instrumental in advancing automated manufacturing processes, particularly in domains requiring high-precision expert robotic motions, such as spray painting and welding. To realize effective automation, robust algorithms are essential for generating extended, object-aware trajectories across intricate 3D geometries. However, contemporary OCMG techniques are either based on ad-hoc heuristics or employ learning-based pipelines that are still reliant on sensitive post-processing steps to generate executable paths. We introduce FoldPath, a novel, end-to-end, neural field based method for OCMG. Unlike prior deep learning approaches that predict discrete sequences of end-effector waypoints, FoldPath learns the robot motion as a continuous function, thus implicitly encoding smooth output paths. This paradigm shift eliminates the need for brittle post-processing steps that concatenate and order the predicted discrete waypoints. Particularly, our approach demonstrates superior predictive performance compared to recently proposed learning-based methods, and attains generalization capabilities even in real industrial settings, where only a limited amount of 70 expert samples are provided. We validate FoldPath through comprehensive experiments in a realistic simulation environment and introduce new, rigorous metrics designed to comprehensively evaluate long-horizon robotic paths, thus advancing the OCMG task towards practical maturity.

Problem

• Generating continuous robot motion paths for complex 3D objects
• Eliminating sensitive post-processing steps in motion generation
• Achieving generalization with limited expert demonstration data

Innovation

• Neural field based end-to-end motion generation
• Continuous function encoding smooth output paths
• Eliminates need for brittle post-processing steps