A Cross-Layer Coding for Scalable ECG streaming

Mobile electrocardiogram (ECG) streaming in body area networks (BANs) is challenging owing to an inherently inconsistent wireless channel, which generally cannot be assumed wide-sense-stationary. Common conventional ECG compression is entropy-based and thus is fundamentally at odds with a BAN channel plagued with variabilities. That is, if the wireless signal experiences a deep fade regime, excessive errors, user contention, and RF interference could all combine so as to result in an interruption in ECG streaming until channel quality recovers. To mitigate against this hard limit on channel quality (i.e. the cliff effect), this paper proposes a linear ECG coding method whereby proneness to mis-reception due to channel errors, contention, and/or interference is traded for a soft, proportional degradation in signal definition. As such, the likelihood of ECG streaming interruption in BANs is vastly lessened while also enhancing capacity and relieving wireless medium contention. This improved robustness and scalability in the wireless network is particularly sought after in mission-critical healthcare applications with stringent QoS demands.