论文部分内容阅读
冲激超宽带生命探测雷达中,传统的平均相消法不能有效去除直达波,从而影响目标回波信号的提取;同时对回波数据中呼吸、心跳信息的提取仍然基于FFT从频域上进行,因而呼吸信号谐波对心跳信号检测的干扰未得到很好解决。针对这两个问题,采用宽带互相关法对回波时延进行估计,然后提取各个回波时延序列的均值与方差作为特征量,采用C-均值聚类算法对回波进行分类,实现直达波与目标回波的分离;采用经验模态分解将目标回波时延序列分解成有限个固有模态函数,从时域上提取呼吸和心跳信号,从而避免呼吸谐波对心跳信号检测的干扰。
Impulse ultra-wideband life detection radar, the traditional average cancellation method can not effectively remove the direct wave, thus affecting the target echo signal extraction; echo data echo, heart beat information extraction is still FFT-based from the frequency domain , So the interference of respiratory signal harmonics on heartbeat signal detection has not been well solved. In view of these two problems, the broadband cross-correlation method is used to estimate the echo delay. Then, the mean and variance of each echo delay sequence are extracted as the feature variables, and the echoes are classified by C-means clustering algorithm to achieve direct Wave and target echoes. Empirical mode decomposition decomposes the target echo delay sequence into a finite number of eigenmodes and extracts the breath and heartbeat signals from the time domain so as to avoid interference of the respiratory harmonics on heartbeat signal detection .