Adaptive ECG compression scheme with prior knowledge support based on compressive sensing

Mobile electrocardiogram (ECG) monitoring systems have sprung up owing to the considerable interest attracted to wireless body area networks (WBAN). The long-term acquisition process for ECG produces large amount of data, which puts forward high demand on sensor lifetime. Fortunately, compressive sensing (CS) theory has been proved useful in energy saving by compressing signal in certain degree and fulfilling transmission. However, the reconstruction error will increase with fixed compression ratio since users or the sparsity of ECG signal will change during monitoring process. This paper concerns the flexibility and reconstruction quality problem existed in traditional CS-based ECG signal processing. One adaptive ECG compression scheme inspired by closed-loop control theory is proposed, in which the compression ratio can be adjusted according to both real-time reconstruction error and prior knowledge support. Simulation results show that the proposed scheme can improve the compression performance of 10.83% compared with traditional CS-based methods.