A Fast and Efficient FMCW Radar Phase Extraction Technique at Ultra-Narrow Range for Vital Sign Detection

A rapid and efficient method for FMCW radar phase extraction is proposed in this paper. This method leverages the phase difference characteristics of FMCW intermediate frequency signals in the fast time dimension and constructs a weight vector to achieve phase acquisition over an ultra-narrow range in the distance dimension. The weight vector is optimized using genetic algorithm to achieve the narrowest possible main-lobe width. The optimized weight vector is stored in an array, allowing for direct addition operations during use. The proposed method surpasses traditional Range-FFT methods in both distance resolution and time complexity. A custom FMCW radar prototype was developed, and experiments were conducted to validate the performance of the proposed method. Results demonstrate that under sleep conditions, the proposed method can effectively detect detailed information about human heartbeats while avoiding interference from abdominal breathing. Compared to ECG, the method achieves an RR interval RMSE of 12ms and heart rate estimation RMSE of 1.3 bpm, providing a foundation for further radar-based biomedical research.