Aiming at the the problem that the array convolutive mixing model of wideband signal in electronic reconnaissance cannot be solved by existing methods, a wideband blind source separation(BSS) algorithm based on frequency invariable transformation(FIT) was proposed in this paper.Firstly, the convolutive mixing problem was transformed into an instantaneous one by forming a frequency invariable network and preprocessing the received signals,and the complexity of convolutive or the frequency-domain BSS algorithms were avoided.Secondly, aiming at the characteristic that wideband signals had strong correlation,a method based on the fractional Fourier transform(FRFT) was proposed to separate the mixing signals, the parameters were estimated and then the signals were recovered according to the strength of the signals, which insured the reliability of separation of weak signals.The simulations indicated that the proposed method had a good robustness and separation effect for signals coincide in time-domain and frequency-domain.%针对已有信号分离算法对电子侦察中的宽带信号阵列接收混合模型不适用的问题,提出了基于频率恒定变换(FIT)的宽带信号分离算法.该算法通过构造频率恒定波束网络对信号进行预处理,将卷积混合问题转换为瞬时混合问题,避免了卷积混合盲源分离和频域盲源分离算法中的计算复杂度;然后针对电子侦察面临的宽带信号相关性强的特点,采用分数阶傅里叶变换,按照信号分量强度依次对混合信号进行参数估计和信号恢复,保证了弱信号分离的可靠性.仿真实验表明,对于时域、频域存在重合的信号,该算法能够实现盲分离,具有良好的鲁棒性.
展开▼
