Towards Safe Imitation Learning via Potential Field-Guided Flow Matching

2025-08-12

AI Summary

In imitation learning, deep generative models—such as flow matching—often lack motion safety in cluttered or obstacle-rich environments. To address this, we propose PF²MP, a novel framework that jointly leverages successful demonstrations to learn both task-specific policies and implicit obstacle potential fields. By dynamically modulating the flow matching generation process via the learned potential field, PF²MP ensures collision-free and efficient task execution. Our method unifies flow matching with artificial potential field principles within an end-to-end trainable architecture, jointly optimizing both the policy network and the implicit potential field representation. It supports control in both task space and joint space. Extensive evaluations on simulated and real-world robotic platforms—including navigation and manipulator manipulation tasks—demonstrate that PF²MP significantly reduces collision rates while maintaining high task success rates, thereby achieving a principled balance between safety and functional performance.

Technology Category

• Machine Learning: Imitation Learning & Inverse Reinforcement Learning
• Intelligent Robots: Motion and Path Planning
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Deep generative models, particularly diffusion and flow matching models, have recently shown remarkable potential in learning complex policies through imitation learning. However, the safety of generated motions remains overlooked, particularly in complex environments with inherent obstacles. In this work, we address this critical gap by proposing Potential Field-Guided Flow Matching Policy (PF2MP), a novel approach that simultaneously learns task policies and extracts obstacle-related information, represented as a potential field, from the same set of successful demonstrations. During inference, PF2MP modulates the flow matching vector field via the learned potential field, enabling safe motion generation. By leveraging these complementary fields, our approach achieves improved safety without compromising task success across diverse environments, such as navigation tasks and robotic manipulation scenarios. We evaluate PF2MP in both simulation and real-world settings, demonstrating its effectiveness in task space and joint space control. Experimental results demonstrate that PF2MP enhances safety, achieving a significant reduction of collisions compared to baseline policies. This work paves the way for safer motion generation in unstructured and obstaclerich environments.

Problem

• Ensuring safe motion generation in imitation learning
• Addressing overlooked safety in complex obstacle environments
• Reducing collisions without compromising task success

Innovation

• Potential Field-Guided Flow Matching Policy
• Modulates flow matching via learned potential field
• Extracts obstacle information from successful demonstrations