一种LFMCW雷达多动目标检测方法

A Novel Method for Multi-Moving Target Detection in LFMCW Radar

为适应车辆防撞及辅助智能驾驶应用需求,针对线性调频连续波雷达因固有距离-速度模糊,在检测多个动目标过程中常出现大量虚假目标,从而造成目标误检率高的问题,提出了一种应用频谱局部相似度的多目标频谱配准方法。在深入研究多个动目标上下扫频段内回波基带信号频谱特征的基础上,该方法通过计算目标上扫频段和下扫频段频谱的相关系数,确定了目标的频谱局部相似度,再依据相似程度进行目标频谱配准,从而去除虚假目标。Matlab环境下的大量仿真实验结果表明:与典型的调频连续波匹配方法相比,该方法有效降低了多个动目标情况下的雷达误检率;在一定速度范围内,误检率为0。

A large number of false targets are often found in the process of detecting multi-moving targets due to the inherent distance-speed blur of LFMCW radar,which often leads to the high false detection rate.In order to meet the requirements of vehicle collision and auxiliary intelligent driving,a multi-moving targets spectrum registration method based on the similarity of local spectrum is presented.After detailed analysis of the spectral characteristics of the echo baseband signal in the upper and lower sweep bands of multiple moving targets,the local similarity of the spectrum is determined by calculating the correlation coefficient of the target s upper and the under sweep frequency spectrums.Then,the false target is removed according to the similarity of target spectrum distribution.The results of a large number of simulation experiments in Matlab show that the proposed method can effectively reduce the radar false detection rate in the case of multi-moving targets and the error rate is zero in a certain speed range compared with the typical FMCW matching method.