mmWave-Diffusion:A Novel Framework for Respiration Sensing Using Observation-Anchored Conditional Diffusion Model

2026-03-21

AI Summary

This work addresses the challenge of non-stationary interference—such as subtle body movements—that degrades reconstruction accuracy in contactless respiration monitoring using millimeter-wave radar. To mitigate this, the authors propose Radar Diffusion Transformer (RDT), an observation-anchored conditional diffusion framework that models the residual between radar phase measurements and ground-truth respiration signals. By initializing sampling within an observation-consistent neighborhood, RDT effectively suppresses interference and enhances denoising efficiency. The method innovatively integrates physical radar observations into a diffusion model through patch-level dual positional encoding, banded-masked multi-head cross-attention, and an observation-anchored sampling strategy. With only 20 reverse diffusion steps, RDT achieves state-of-the-art performance in both respiration waveform reconstruction and respiratory rate estimation, significantly reducing inference cost while improving generalization.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Planning, Routing, and Scheduling: Activity and Plan Recognition
• Intelligent Robots: State Estimation

Application Category

• Responsible Web: Data and user privacy-enhancing technologies for the Web
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Millimeter-wave (mmWave) radar enables contactless respiratory sensing,yet fine-grained monitoring is often degraded by nonstationary interference from body micromotions.To achieve micromotion interference removal,we propose mmWave-Diffusion,an observation-anchored conditional diffusion framework that directly models the residual between radar phase observations and the respiratory ground truth,and initializes sampling within an observation-consistent neighborhood rather than from Gaussian noise-thereby aligning the generative process with the measurement physics and reducing inference overhead. The accompanying Radar Diffusion Transformer (RDT) is explicitly conditioned on phase observations, enforces strict one-to-one temporal alignment via patch-level dual positional encodings, and injects local physical priors through banded-mask multi-head cross-attention, enabling robust denoising and interference removal in just 20 reverse steps. Evaluated on 13.25 hours of synchronized radar-respiration data, mmWave-Diffusion achieves state-of-the-art waveform reconstruction and respiratory-rate estimation with strong generalization. Code repository:https://github.com/goodluckyongw/mmWave-Diffusion.

Problem

• mmWave radar
• respiration sensing
• micromotion interference
• nonstationary interference
• contactless monitoring

Innovation

• conditional diffusion model
• mmWave radar
• respiration sensing
• observation-anchored initialization
• Radar Diffusion Transformer