Power-efficient joint compressed sensing of multi-lead ECG signals

Abstract

Abstract-Compressed Sensing (CS) is a new acquisitioncompression paradigm for low-complexity energy-aware sensing and compression. By merging both sampling and compression, CS is very promising to develop practical ultra-low power readout systems for wireless bio-signal monitoring devices, where large amounts of sensor data need to be transferred through power-hungry wireless links.Lately CS has been successfully applied for real-time energyaware single-lead ECG compression on resource-constrained Wireless Body Sensor Network (WBSN) motes [1]. Building on our previous work, in this paper we propose a new and promising approach for joint compression of multi-lead ECG signals, where strong correlations exist between them. This situation that exhibit strong correlations, can be exploited to reduce even further amount of data to be transmitted wirelessly, thus addressing the important challenge of ultra-low-power embedded monitoring of multi-lead ECG signals.I. INTRODUCTION CS is a new sensing and processing paradigm, which challenges the traditional analog-to-digital conversion based on the Shannon sampling theorem. For sparse signals such as the electrocardiogram (ECG), Nyquist-rate sampling produces a large amount of redundant digital samples, which are costly to wirelessly transmit in the context of our target mobile ECG monitoring systems, and require to be further compressed using non-linear digital techniques. CS is a methodology that has been recently proposed to address this problem.Capitalizing on this sparsity, we have recently proposed [1], [2], to apply the emerging compressed sensing (CS) approach [3] for a low-complexity, real-time and energy-efficient ECG signal compression on WBSN motes. We have also quantified the potential of compressed sensing (CS) for lowcomplexity energy-efficient ECG compression on the stateof-the-art Shimmer TM WBSN mote. Interestingly, our results show that CS represents a competitive alternative to state-ofthe-art digital wavelet transform (DWT)-based ECG compression solutions in the context of WBSN-based ECG monitoring systems. The results validate the suitability of compressed sensing for real-time energy-aware ECG compression on resource-constrained WBSN motes.Nonetheless, in a real scenario many of the bio-signals are acquired on multiple channels. In these cases, sampling, compressing and reconstructing each signal individually is clearly sub-optimal, because leads are not independent sources, and are in fact strongly correlated. In the case of the ECG acquisitions, all signals can be considered as different projections of a single multidimensional source, which is the electrical

Ultra-Low Power Design of Wearable Cardiac Monitoring Systems

“…For this application, the measured sensitivity and specificity of retrieved fiducial points are above 90% in all cases, which is at the target level for medical use in this field. Figure 5 compares the averaged signal-to-noise ratio (SNR) results over different compression ratios (CR) for singlelead and multi-lead CS compression [6]. These results show that an averaged SNR over 20 dB (corresponding to good reconstruction quality [16]) is reached for CR = 65.9% and CR = 72.7% for single and multi-lead CS, respectively.…”

“…Second, in case of the multi-lead ECG compression, there is a strong correlation between the sparsity structure among the leads, each lead therefore conveying useful information about other leads. In particular, non-zero coefficients are partitioned in subsets or groups, and this information can be employed to enhance the compression performance across all leads [6].…”

Energy vs. Reliability Trade-offs Exploration in Biomedical Ultra-Low Power Devices

“…This process enables a relaxation of the traditional reliability requirements (i.e., 100% computational precision is not needed), which can be exploited by the CS application. In particular, the maximum required output SNR to get almost a 100% reconstruction quality is only 35 dB in the case of multi-lead ECG [10] and 40 dB in the case of a single lead ECG [11]. Thus, as Fig.…”

“…It helps to reduce airtime over energy-hungry wireless links. We have implemented our own version of the algorithm presented by the authors of [10], which takes as input a vector of ECG samples and applies a 50% lossy compression algorithm to convert it into a smaller one (half the size of the input vector).…”

Approximate compressed sensing

“…In such a situation, where non-zero coefficients are naturally partitioned in subsets or groups, the best choice could be using a group-sparsity inducing term [22]. In a recent prior work [21], we proposed to replace the ℓ1 norm with mixed ℓ1/ℓ2 norm. It behaves like an ℓ1-norm on the vector ( αi 2 )i∈L in R |L| , and therefore, induces group sparsity.…”