| 摘要: |
| 应用雷达进行生命体征检测时，为解决传统算法无法区分相同距离的多个目标和多目标生命体征信号检测易受干扰的问题，本文提出一种新型的基于调频连续波（FMCW）雷达的多目标生命体征检测与干扰抑制算法。基于MIMO雷达回波数据进行距离-方位维快速傅里叶变换（FFT），完成距离和方位角的目标定位，解决同距离目标分辨问题，之后利用方差法滤除静态杂波和直流分量。针对生命体征信号易受干扰的问题，本文算法利用变分模态分解（VMD）进行呼吸和心跳信号的分离与重构，剔除无效的信号分量，然后对有效重构信号进行自相关和奇异值分解（SVD）去噪，得到干净的呼吸和心跳信号。通过多组实验测试，本算法相较于传统的FFT频谱分析法、VMD信号重构法和基于完全噪声辅助聚合经验模态分解联合独立成分分析算法（CEEMDAN-ICA），心率检测的平均误差分别降低了8.43%、5.66%、1.43%，验证了所提算法的有效性。此外，本文还进行了算法鲁棒性实验和工程应用边界条件分析。 |
| 关键词: 调频连续波雷达 生命体征检测 自相关 奇异值分解 变分模态分解 |
| DOI： |
| 分类号:TN957.51 |
| 基金项目:国家自然科学基金项目（面上项目，重点项目，重大项目） |
|
| Research on multi-target non-contact vital signs detection algorithm based on FMCW radar |
|
|
|
|
| Abstract: |
| In order to solve the problem that traditional algorithms cannot distinguish between multiple targets and multi-targets at the same distance when using radar for vital signs detection, a novel multi-target vital signs detection algorithm based on frequency modulate continuous wave (FMCW) radar is proposed in this paper. Based on MIMO radar echo data, the range-azimuth Fast Fourier Transform (FFT) is used to locate the range and azimuth angle of the target, and to solve the problem of the same range target resolution. Then the variance method is used to filter the static clutter and the DC component. Aiming at the problem that vital signs signals are easily disturbed, this paper uses Variational Mode Decomposition (VMD) algorithm to separate and reconstruct breathing and heartbeat signals, eliminate invalid signal components. Then, the reconstructed signals are denoised by Autocorrelation and Singular Value Decomposition (SVD) to obtain clean breathing and heartbeat signals. Through multiple experimental tests, compared with traditional FFT spectrum analysis method, VMD signal reconstruction method and Complete Ensemble Empirical Mode Decomposition with Adaptive Noise Combined Independent Component Analysis Algorithm（CEEMDAN-ICA）, the average error of heart rate detection of this algorithm is reduced by 8.43%, 5.66% and 1.43%, respectively, which verifies the effectiveness of the proposed algorithm. In addition, the robustness experiment of the algorithm and the analysis of the boundary conditions of engineering applications are also carried out. |
| Key words: FMCW radar vital signs detection Autocorrelation Singular Value Decomposition Variational Mode Decomposition |