A Power-Aware ECG Transmission Framework with Server Aided Lossless Compression

Wearable sensor nodes based WBAN system is utilized to reduce individuals' life risk by detecting various cardiac anomalies via remote ECG signal monitoring. In this context, a power-aware WBAN transmission system through a server-aided ECG compression technique is presented in this paper. For that, a lossless compression technique to deal with the power consumption issue of a sensor node is proposed. The proposed compression approach employs a frame-adaptive Golomb-rice coding in coordination with k-means clustering at the remote server. The proposed algorithm effectively achieves a similar compression ratio under different levels of noise incorporated in the digitized ECG signal. The algorithm is validated with ECG signals from the MIT-BIH arrhythmia database, resulting in an average compression ratio of 2.89. The VLSI architecture of the proposed technique is implemented on TSMC 90 nm technology, which consumes a power of 65
$\mu W$
with 0.0049
$mm^{2}$
area overhead.