WiRM: Wireless Respiration Monitoring Using Conjugate Multiple Channel State Information and Fast Iterative Filtering in Wi-Fi Systems

2025-07-31

AI Summary

This paper addresses severe waveform distortion and insufficient noise robustness in Wi-Fi-based non-contact respiration monitoring. To this end, we propose an end-to-end wireless respiratory sensing framework. Methodologically, we introduce two novel techniques: Conjugate Multiplicative Phase Purification (CMPP) and Adaptive Multi-Trajectory Carving (AMTC), integrated with respiration-rate-guided waveform decomposition and fast iterative filtering to achieve high-fidelity reconstruction of CSI phase signals. Experimental results demonstrate that our approach reduces the root-mean-square error in respiration rate estimation by 38% on average; improves the mean absolute correlation between reconstructed and ground-truth respiratory waveforms by 178.3%; and maintains superior robustness against thermal noise, multiplicative noise, and phase noise. To the best of our knowledge, this is the first work to jointly incorporate CMPP and AMTC for respiratory waveform recovery, significantly enhancing both accuracy and practicality of wireless physiological sensing.

Technology Category

• Planning, Routing, and Scheduling: Activity and Plan Recognition
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Computer Vision: Multi-modal Vision

Application Category

• Responsible Web: Measurement, analysis, and circumvention of Web censorship
• Web Mining and Content Analysis: Web measurements
• Systems and Infrastructure for Web, Mobile and WoT: Data management and stream processing for Web, mobile and wireless applications

Abstract

Monitoring respiratory health with the use of channel state information (CSI) has shown promising results. Many existing methods focus on monitoring only the respiratory rate, while others focus on monitoring the motion of the chest as a patient breathes, which is referred to as the respiratory waveform. This paper presents WiRM, a two-staged approach to contactless respiration monitoring. In the first stage, WiRM improves upon existing respiratory rate estimation techniques by using conjugate multiplication for phase sanitisation and adaptive multi-trace carving (AMTC) for tracing how the respiratory rate changes over time. When compared against three state-of-the-art methods, WiRM has achieved an average reduction of $38%$ in respiratory rate root mean squared error (RMSE). In the second stage, WiRM uses this improved respiratory rate estimate to inform the decomposition and selection of the respiratory waveform from the CSI data. Remarkably, WiRM delivers a $178.3%$ improvement in average absolute correlation with the ground truth respiratory waveform. Within the literature, it is difficult to compare the robustness of existing algorithms in noisy environments. In this paper, we develop a purpose-built simulation toolkit to evaluate the robustness of respiration monitoring solutions under various noise conditions, including thermal, multiplicative, and phase noise. Our results show that WiRM demonstrates improved or comparable resilience to these common noise sources.

Problem

• Improves respiratory rate estimation using conjugate multiplication and AMTC
• Enhances respiratory waveform extraction from CSI data
• Evaluates robustness in noisy environments with simulation toolkit

Innovation

• Conjugate multiplication for phase sanitisation
• Adaptive multi-trace carving for rate tracing
• Simulation toolkit for noise resilience evaluation