A Bayesian Approach for Estimating Heart Rate using FMCW Radar

Abstract

This paper illustrates a Markov model of peak-finding and gives a Bayesian method for estimating heart rate based on an FMCW radar. The Bayesian method provides a framework for estimation and inference and noise peaks can be filtered out. Conjugate priors are used making the algorithm computation-ef-ficient. A real-time algorithm is given and several experiments have been carried out. It is shown that the accuracy of the heart rate is close to reference and much more stable than the traditional FFT method and promising in overcoming noise.