Battery-aware approximate wireless telemetry framework for resilient wearable ECG monitoring under extreme power constraints.

Wearable electrocardiogram (ECG) patches utilizing Bluetooth Low Energy (BLE) face a critical, yet under- characterized, failure mode: as battery state of charge (SoC) depletes, firmware-mandated reductions in RF transmission power elevate the bit error rate (BER) in Rayleigh-fading channels, causing conventional QRS detection to fail during the most clinically important periods of continuous cardiac monitoring. This paper presents the Battery-Aware Approximate Wireless Telemetry (BAWT) framework, a co-designed solution that jointly optimizes a Peukert-corrected Li-ion discharge model, a six-state adaptive RF power controller, and a Rayleigh-fading indoor channel model. At the receiver, the proposed Bayesian Adaptive Feature Estimator (BAFE) fuses a Wiener-optimal morphological bandpass prior with a channel-SNR-derived MMSE-Wiener weight, enabling reliable QRS extraction from severely corrupted bit streams without forward error correction hardware. Validated on the MIT-BIH Arrhythmia and PTB-XL clinical databases against the ANSI/AAMI EC57 standard, BAWT demonstrates substantial performance gains: at the clinically critical 20% SoC operating point, BAFE raises mean QRS sensitivity from 35.1% to 91.9% on MIT-BIH and from 32.9% to 82.8% on PTB-XL, with the majority of individual records satisfying the Se ≥ 95% clinical threshold. The adaptive power controller extends the critical operating window by 355% over fixed full-power operation, and a fully characterized Pareto-optimal frontier enables system designers to navigate the trade-off between battery lifetime extension (up to 144.5%) and clinical QRS detection accuracy across the full SoC range. These results establish a rigorous co-design framework for robust, power-aware wearable cardiac monitoring compliant with ANSI/AAMI EC57. .