Contactless Heartbeat Estimation with FMCW Radar

Measuring heart rate is usually done with the use of invasive technologies, such as an electrocardiogram which uses electrodes. However, this may not always be the optimal method. Non-invasive methods for measuring heart rate are often preferred due to their simplicity and reduced discomfort for the individual. Using Millimetre-wave frequency modulated continuous wave radars (FMCW) to detect chest movements and extract vital signs is a promising possible alternative due to its non-invasive nature. In this paper we use Fast Fourier transformations, the Savitzky–Golay filter, Elliptic filters, and the Kalman Filter to explore the feasibility of real-time contactless heart rate estimation using FMCW radar. Across all datasets, using a 2.5s monitoring window, an average Mean Absolute Percentage Error (MAPE) of 5.56% was achieved. Remarkably, on an individual dataset, the MAPE reached an impressive low of 3.41%, highlighting the robustness of the approach in advancing real-time contactless heart rate estimation with FMCW radar.