Human Presence Detection Using Ultrashort-Range FMCW Radar Based on DCNN

This study introduces a methodology for detecting human presence in close proximity using frequency-modulated continuous wave radar. We focus on discerning human presence against inanimate objects by analyzing target vibrations. Instead of relying solely on conventional features, such as magnitude and phase variances of the received signal, we propose utilizing phasor scatter plots and spectrograms for capturing statistical and time-varying features. To process these two 2-D features, this research suggests the integration of deep convolutional neural networks (DCNNs), followed by deep neural networks (DNNs) for feature fusion, thus enhancing target classification accuracy. Through a comprehensive measurement campaign, we demonstrate the performance of the proposed methodology for human detection in close proximity. While the performance of conventional machine learning methods is below 95%, the result of our proposed method for human detection stands at 99.83%, indicating its potential to contribute to the efficient control of transmission power of smart devices, ensuring compliance with regulations on electromagnetic (EM) radiation.