减法聚类的多分量线性调频信号检测与参数估计算法

针对Wigner-Hough变换在低信噪比下易产生伪峰,导致在进行检测时阈值选取困难,无法准确对多分量线性调频信号进行检测和参数估计的问题,提出了一种基于减法聚类的平滑伪Wigner-Hough(SPWHT)变换的检测和参数估计算法。首先采用平滑伪Wigner-Hough变换(SPWHT),通过平滑交叉项和噪声来达到抑制干扰的目的,并运用阈值分割的图像处理方法进一步消噪,减少Hough变换后所产生的伪峰;再利用较低门限对Hough变换后的图像进行峰值检测,防止对信号的漏检。最后,对于备选峰值进行减法聚类的二次处理,实现信号参数寻优,去除伪峰影响和时频分辨率下降引起的峰值展宽,完成对多分量信号的准确检测和参数估计。该算法通过聚类避免了阈值设置的困难,解决了低信噪比下多分量线性调频信号的检测与参数估计问题。仿真结果表明了在-10 d B信噪比下该算法仍有较好的检测及估计性能。

Detection and parameter estimation of multi-component LFM signals algorithm based on subtractive clustering

To solve the problem of pseudo peak interference and threshold choise in radar signals detection and parameter estimation using Wigner-Hough transform, a detection and estimation algorithm with Smooth Pseudo Wigner-Hough transform(SPWHT) based on subtractive clustering is proposed. Firstly ,the time-frequency distribution image of multi-component LFM signals is obtained by the SPWVD to reduce the interference of noise and cross-term ,so that the pseudo peak will be less after the Hough transform. Then choose a low threshold to detect the peak in SPWHT image, avoiding lose signal. Finally, use subtractive clustering to process the optional peaks and clean the interference of pseudo peak. After that, all the LFM signals are detected by the algorithm and the parameters of them can be estimated. The algorithm avoid the difficulty of setting the threshold ,solving the problem of the detection and parameter estimation of multi-component LFM signals under the low SNR. Simulation results show the efficacy of the proposed algorithm.