Constraint-Aware Diffusion Guidance for Robotics: Real-Time Obstacle Avoidance for Autonomous Racing

2025-05-19

AI Summary

To address the lack of constraint-awareness and physical feasibility guarantees in diffusion models for safety-critical robotic tasks—such as real-time autonomous racing—this paper proposes the Constraint-aware Diffusion Guidance (CoDiG) framework. CoDiG explicitly incorporates barrier functions into the diffusion denoising process, enabling joint enforcement of kinematic and obstacle-avoidance constraints. It further integrates real-time diffusion sampling optimization with a lightweight deployment architecture, validated end-to-end on a miniature racing platform. The method achieves strong generalization to unseen tracks and dynamic obstacle configurations using only minimal training data. Experimental results demonstrate a 72% reduction in collision rate and inference latency under 15 ms, significantly improving both safety and real-time performance. CoDiG overcomes the fundamental limitation of conventional diffusion models in satisfying hard physical constraints, establishing a new paradigm for deploying diffusion-based control in safety-critical robotics.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Humans and AI: Human-Aware Planning and Behavior Prediction

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: Fairness-aware retrieval and ranking

Abstract

Diffusion models hold great potential in robotics due to their ability to capture complex, high-dimensional data distributions. However, their lack of constraint-awareness limits their deployment in safety-critical applications. We propose Constraint-Aware Diffusion Guidance (CoDiG), a data-efficient and general-purpose framework that integrates barrier functions into the denoising process, guiding diffusion sampling toward constraint-satisfying outputs. CoDiG enables constraint satisfaction even with limited training data and generalizes across tasks. We evaluate our framework in the challenging setting of miniature autonomous racing, where real-time obstacle avoidance is essential. Real-world experiments show that CoDiG generates safe outputs efficiently under dynamic conditions, highlighting its potential for broader robotic applications. A demonstration video is available at https://youtu.be/KNYsTdtdxOU.

Problem

• Enabling real-time obstacle avoidance in autonomous racing
• Integrating constraint-awareness into diffusion models for robotics
• Ensuring safety with limited training data in dynamic environments

Innovation

• Integrates barrier functions into denoising process
• Ensures constraint satisfaction with limited data
• Real-time obstacle avoidance for autonomous racing