A Data-Driven Compressive Sensing Framework for Long-Term Health Monitoring

TL;DR

This paper introduces a data-driven compressive sensing framework that learns from patient data to improve long-term health monitoring, achieving higher compression ratios and noise resilience for ECG signals.

Contribution

It presents a novel co-training approach that optimizes sensing matrices and dictionaries based on patient data, enhancing CS performance and noise robustness.

Findings

Achieves up to 80% higher compression ratio compared to traditional methods.

Demonstrates tolerance to 40dB higher noise energy at a compression ratio of 9.

Improves reconstruction quality of ECG signals significantly.

Abstract

Compressive sensing (CS) is a promising technology for realizing energy-efficient wireless sensors for long-term health monitoring. In this paper, we propose a data-driven CS framework that learns signal characteristics and individual variability from patients' data to significantly enhance CS performance and noise resilience. This is accomplished by a co-training approach that optimizes both the sensing matrix and dictionary towards improved restricted isometry property (RIP) and signal sparsity, respectively. Experimental results upon ECG signals show that our framework is able to achieve better reconstruction quality with up to 80% higher compression ratio (CP) than conventional frameworks based on random sensing matrices and overcomplete bases. In addition, our framework shows great noise resilience capability, which tolerates up to 40dB higher noise energy at a CP of 9 times.