与常规雷达相比,超宽带雷达具有距离分辨力高、近距离盲区小、穿透性强、目标识别率高等特点,已被广泛应用于灾后搜寻、救援工作中,以对受困生命体征目标进行生命探测。为实现使用超宽带雷达对受困生命体征目标的识别定位,本研究提出基于信号多特征提取技术及支持向量机模型的人体呼吸信号识别方法。首先,使用经验模态分解、变分模态分解及希尔伯特变换提取雷达探测信号的微多普勒特征,使用傅里叶变换提取宏观频谱特征,使用相关分析获取相关性特征;然后,以提取的信号特征为输入,使用支持向量机模型对信号进行分类,进而对人体呼吸信号进行识别,对人体位置进行定位。不同障碍物场景下的试验结果表明,本方法可有效识别砖墙、建筑楼板等遮挡物下的受困生命体征目标,并提供其位置信息。

Compared with conventional radar, the ultra-wideband (UWB) radar has the advantages of high distance resolution, small blind spot at close range, strong penetration, and high target recognition rate. Therefore, it has been widely used in post-disaster Radar-based Life Detecting System. In order to identify and locate the trapped person using UWB radar, a method based on signal multi-feature extraction and support vector machine model for human respiratory signal recognition is proposed in this paper. Firstly, we use empirical mode decomposition, variational mode decomposition, and Hilbert transformation to extract the micro-Doppler characteristics of echo signals use Fourier transformation to extract the spectrum characteristics and use correlation analysis to obtain the correlation characteristics. Then, the signals are classified by a support vector machine model based on these signal features. As a result, the respiratory signal can be identified and the position of the human body can be located. The experimental results obtained from different scenes show that the proposed method can effectively identify the human body which is shielded by brick walls and floor slabs, and the location of the human body can be determined at the same time.