Indoor Personnel Detection Method Based on Millimeter-Wave Radar

In recent years, technological advancements have paved the way for millimeter-wave radar to exhibit extensive potential applications across diverse fields, including smart homes, smart elderly care, and smart offices. This paper endeavors to leverage the two-dimensional heatmap data derived from Frequency-Modulated Continuous Wave (FMCW) millimeter-wave radar and implement YOLOv7 deep learning algorithm, which is grounded in Convolutional Neural Networks (CNN), for indoor personnel detection. Initially, the paper delves into the design of an indoor personnel detection system based on millimeter-wave radar. By amassing data on indoor personnel distribution and activities in authentic environments, we have crafted heatmaps that depict indoor personnel distribution. Furthermore, the paper proposes a fusion of YOLOv7 with radar signal processing techniques. Subsequently, the YOLOv7 algorithm is harnessed to create a visual representation of indoor personnel positions. To validate the indoor personnel detection system on the basis of millimeter-wave radar put forward in this paper, rigorous functional tests were conducted indoors. The findings reveal that this methodology can swiftly detect indoor personnel with a remarkable accuracy rate of 99.50%, underscoring its robustness and effectiveness.