基于时频独立分量分析的Wigner-Ville分布交叉项消除法

由于信号混合引起的交叉项严重降低了Wigner-Ville分布的时频分辨率，为此提出一种基于独立分量分析的Wigner-Ville分布交叉干扰项消除法.在无须了解信号混合系数的情况下，通过盲源分离法提取各独立分量信号，给出盲源分离结果不确定的解决方法.建立了包含自项与交叉项的时频分布矩阵，利用时频分布矩阵的联合对角化算法消除独立分量信号之间的交叉干扰项.通过独立分量自项求和重构Wigner-Ville分布，重构的Wigner-Ville分布具有良好的时频分辨率.数值仿真结果验证了算法的有效性.

Time frequency-based independent component analysis for elimination of cross-terms in Wigner-Ville distribution

The cross-terms of Wigner-Ville distribution resulting from signals mixing substantially reduce its time-frequency resolution. An independent component analysis approach was presented to eliminate the cross-terms in Wigner-Ville distribution. The ambiguity of separated components was firstly resolved and the independent signal components were extracted via blind source separation. The time-frequency distribution matrices containing auto-terms and cross-terms were established. Based on joint diagonalisation of the time-frequency distribution matrices, the proposed independent component analysis method recovered the auto-term of each component in Wigner-Ville distribution. The cross-terms caused by independent component mixing were reduced significantly. The reconstructed Wigner-Ville distribution has good time-frequency resolution. The method was finally illustrated by a numerical example.