基于FastICA算法和小波变换的雷达信号分选

Sorting of radar signals based on FastICA algorithm and wavelet transfer

传统的独立分量分析(ICA)算法对噪声敏感,存在很难正确分选带噪混合雷达信号的问题。针对该问题提出一种结合FastICA算法和小波去噪的改进算法。该算法首先利用小波阈值法对带噪雷达信号进行去噪,适当提高信噪比后再用FastICA算法进行分离,最后进一步对分离信号作矢量归一和再消噪处理,得到各个雷达源信号的最终估计。仿真结果表明,与传统的ICA算法相比,该改进算法可以有效地去除噪声,提高带噪雷达信号分选的准确率。

It＇s hard to sort the noisy mixed radar signals, for the traditional ICA algorithm is sensitive to noise. Aiming at this problem, an advanced algorithm is proposed, which is in combination with FastlCA algorithm and wavelet denoising. The new algorithm denoises the radar signals with the wavelet threshold method first, then after the SNR increased, the radar signal is separated through FastlCA algorithm, finally the separated signal is post-processed, then the estimated source radar sig- nals are obtained. The simulation result shows that compared with the traditional ICA, the new algorithm can effectively denoise the signal and improve the accuracy of radar signals sorting.