| 摘要: |
| 针对当前使用体征信号进行身份验证准确率低，且特征提取过程复杂的问题，本文在通过毫米波雷达检测生命体征的基础上，提出了一种将纯净的人体胸腔信号（Chest Cavity Signal, CCS）作为样本进行身份验证的方法。首先，对提取到的雷达原始信号进行预处理，消除与实验无关的冗余干扰并提取相位信号。接着对含有干扰的相位信号进行变分模态分解（VMD），提取纯净的心跳与呼吸信号并制作CCS样本。最后将CCS样本送入二维卷积神经网络（2D CNN）中进行训练并验证身份，识别准确率达到了97.5%，实验证明本文提出的方法对于身份验证具有很好的效果。 |
| 关键词: 毫米波雷达 身份验证 变分模态分解 二维卷积神经网络 胸腔信号 |
| DOI：DOI:10.3969/j.issn.1672-2337.2023.05.010 |
| 分类号:TN957 |
| 基金项目: |
|
| FMCW Radar Authentication Based on Chest Cavity Signal Samples |
|
QI Jing， WANG Zhengdong， XIE Guangzhi
|
|
School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, China
|
| Abstract: |
| In order to solve the problem of low accuracy and complex feature extraction process in the current use of sign signals for identity verification, this paper proposes a method to use pure chest cavity signal （CCS） as a sample for identity verification based on the detection of vital signs by millimeter wave radar. First, the raw radar signal is preprocessed to eliminate the redundant interference irrelevant to the experiment and extract the phase signal. Then, variational modal decomposition （VMD） is performed on the phase signal containing interference to extract pure heartbeat and respiration signals and make CCS samples. Finally, CCS samples are sent into 2D convolutional neural network （2D CNN） for training and verification of identity, and the recognition accuracy rate reaches 97.5%. The experiment proves that the proposed method has a good effect on identity verification. |
| Key words: millimeter wave radar identity authentication VMD 2D CNN CCS |
