Compressive Sampling and Reconstruction of ECG Signal for Manned Spaceflight Applications

Electrocardiogram (ECG) is a vital signal which represents the state of health of astronauts in a manned spaceflight mission and hence must be continuously acquired and transmitted throughout the mission. Telemetry bandwidth is a premium resource in such applications. Traditional method of sampling at Nyquist rate is highly bandwidth inefficient. Hence it is advantageous to use Compressive sensing (CS) technique to optimize data at measurement point itself. Conventional CS techniques employ computationally intensive measurement matrices which are not hardware efficient for both acquisition as well as recovery. In this work, a hardware efficient scheme with use of a sparse binary measurement matrix is proposed. The signal is recovered from the compressive measurement using a constraint function based on inverted Laplace distribution function. Gradient decent method is used to recursively recover the original ECG signal from the compressive measurements in an efficient manner. Apart from this, a projection scheme is also proposed to minimize the recovery error. The proposed scheme was extensively evaluated with different ECG samples with different compression ratios. Finally, the proposed scheme was benchmarked with Approximated L0 norm based method and it is found to perform more efficiently in compressively sensing and recovery of ECG signals.