[Paper Review] MIRO: Multi-radar Identity and Ranging for Occupational Safety

Tirthankar Halder, Argha Sen+3 more|arXiv (Cornell University)|Mar 8, 2026

Air Quality Monitoring and Forecasting0 citations

TL;DR

MIRO presents a privacy-preserving framework that uses a network of mmWave radars and localized PM sensors to perform multi-radar worker re-identification and personalized exposure estimation in outdoor/industrial settings. It introduces a Pix2Pix-based view adaptation and a TDSCAN clustering to maintain identity consistency across overlapping radar viewpoints.

ABSTRACT

Occupational exposure to airborne particulate matter (PM) poses a severe health risk in open industrial workspaces such as stonecutting yards. Conventional monitoring solutions such as wearable PM sensors and camera-based tracking are impractical due to discomfort, maintenance issues, and privacy concerns. We present MIRO, a privacy-preserving framework that integrates continuous PM sensing with a multi-radar millimeter-wave (mmWave) re-identification (re-ID) backbone. A distributed network of PM sensors captures localized pollutant concentrations, while spatially overlapping mmWave radars track and re-associate workers across viewpoints without relying on visual cues. To ensure identity consistency across radars, we introduce a GAN-based view adaptation network that compensates for azimuthal distortions in range-Doppler (RD) signatures, combined with correlation-based cross-radar matching. In controlled laboratory experiments, our system achieves a re-ID F1-score of 90.4% and a mean Structural Similarity Index Measure (SSIM) of 0.70 for view adaptation accuracy. Field trials in rural stone-cutting yards further validate the system's robustness, demonstrating reliable worker-specific PM exposure estimation.

Motivation & Objective

• Motivate accurate, fine-grained occupational exposure monitoring that preserves worker privacy and avoids wearable sensors or cameras.
• Develop a multi-radar re-ID system to sustain identity across overlapping mmWave radar views in industrial spaces.
• Introduce a view-adaptation mechanism to normalize azimuth-dependent micro-Doppler signatures across viewpoints.
• Link worker trajectories with localized PM measurements to estimate worker-specific pollution exposure.

Proposed method

• Use a distributed PM sensor network to capture localized pollutant concentrations.
• Deploy multiple mmWave radars with overlapping fields of view for cross-view worker tracking.
• Extract RD heatmaps centered on each worker to capture activity signatures.
• Apply a Pix2Pix-based view adaptation network to translate RD signatures between radar viewpoints.
• Employ TDSCAN for Doppler-aware, temporally-consistent clustering to localize workers across radars.
• Compute cross-radar association scores to propagate identity and form a global identity graph.
• Extract activity signatures and combine them with PM data to estimate worker-specific exposure.

Experimental results

Research questions

• RQ1 Can MIRO reliably re-identify workers across multiple overlapping mmWave radar viewpoints in semi-static industrial settings?
• RQ2 Does azimuth-aware view adaptation normalize micro-Doppler RD signatures across viewpoints while preserving activity structure?
• RQ3 Can cross-radar association sustain identity over time and occlusions to produce coherent worker trajectories?
• RQ4 Is the system able to produce reliable worker-specific PM exposure estimates in real-world stone-cutting environments?

Key findings

• MIRO achieves a re-ID F1-score of 90.4% in lab tests.
• The view-adaptation network achieves a mean SSIM of 0.70 for cross-view RD translation.
• Field deployments validate robustness across open-air, semi-mechanized, and indoor stone-processing environments.
• PM exposure analyses show activity-dependent variations and demonstrate the need for task-aware exposure mapping.
• The system demonstrates privacy-preserving, device-free monitoring suitable for dusty environments.