Comprehensive signal processing approaches for non-contact heartbeat detection using 24 GHz FMCW radar

Abstract. Non-contact monitoring of vital signs has become increasingly important for enhancing patient comfort and reducing interference during medical assessments. Such methods are particularly valuable for addressing phenomena such as white-coat hypertension and for enabling more accurate diagnoses in home environments. This paper consolidates three innovative approaches for heartbeat detection using a dedicated 24 GHz frequency-modulated continuous-wave (FMCW) radar. First, a novel signal envelope analysis method is introduced, demonstrating its ability to extract heartbeat information and to validate the results against an ECG reference system (GE Healthcare CARESCAPE Monitor B650). Second, computationally intensive steps are optimized for real-time processing by replacing the Hilbert transform and smoothing functions with more efficient alternatives such as Moving-RMS and Moving-Average filters. These adaptations enable low-latency operation while maintaining reliable detection performance. Lastly, a convolution-based Gaussian pulse method is presented to extend the envelope analysis, followed by a time-domain plausibility check to refine peak detection. The modular radar system comprises a 2TX/4RX MIMO radar front-end, a filtering and amplification module, and a data acquisition module, all developed at the Laboratory of Applied Radar and Optical Systems (LaROS) at Trier University of Applied Sciences. Each proposed method has been optimized for the 24 GHz ISM band, offering a balance between real-time capability and detection accuracy. Validation against ECG benchmarks confirms the feasibility and effectiveness of the presented approaches, contributing to the advancement of non-contact vital sign monitoring technologies.