Accommodate Data Loss in Monitoring Vital Signs Through Autoregressive Model

In this paper, the measurement of two vital signs including heartbeat and respiratory rate are discussed under two critical scenarios; namely subjection to noise and intermittent observations. Previously adopted scheme for finding the above mentioned vital signs was Fourier Transform which couldn't handle non-stationary process. For a broader perspective, Wavelet Transform is employed in this paper which is equally applicable to stationary and non-stationary processes. In addition, the intermittent observation is a malfunction which may result in severe consequences in measuring vital signs. In past, only noise-free data has been incorporated in tracing vital signs parameters. A Modified Robust Kalman Filter (MRKF) is designed to obtain optimum results in the presence of above two critical scenarios. Simulation results obtained on real data show that the performance of MRKF produces similar vital signs as with clean and undistorted data.