基于改进SO-CFAR和ACA-VMD算法的雷达生命体征检测

非接触生命体征检测技术难以有效利用胸腔产生的微多普勒效应提取心跳和呼吸信号，针对这一问题，文中提出了一种基于改进最小选择恒虚警(SO-CFAR)和蚁群变分模态分解(ACA-VMD)算法的生命体征检测方法，并通过仿真和实测验证了算法的检测精度。首先对77 GHz毫米波雷达的中频回波信号进行预处理得到干净的雷达I/Q数据，然后调整因子以平衡前后窗的功率水平让单元极小值恒虚警检测能够对噪声下的目标进行精确提取，最后采用蚁群优化后的变分模态对目标信号进行模态混叠的抑制并采用全相位频谱分析，使得呼吸和心跳的信噪比改善了1.765 dB,完成呼吸和心跳有效分离和提取，实现了人体生命体征的准确检测。

Radar Vital Signs Detection Based on Improved SO-CFAR and ACA-VMD Algorithms

In this paper, we propose a method for detecting vital signs based on the modified least selected constant false alarm (SO-CFAR) and ant colony variational modal decomposition (ACA-VMD) algorithms, and validate the accuracy of the algorithm through simulations and experiments. The ACA-VMD algorithm is firstly used to obtain clean radar I/ Q data by pre-processing the IF echo signal of 77 GHz millimeter wave radar, and then the power level of the front and rear windows is adjusted to allow the cell-minimum constant false alarm detection to accurately extract the target under noise. The signalto- noise ratio of respiration and heartbeat is improved by 1. 765 dB, and the respiration and heartbeat are effectively separated and extracted to achieve the accurate detection of human vital signs.