非平稳阵列信号盲分离及时变信道均衡的研究

This project has completed several research works on array-signal blind separation and blind channel equalization under support of NSF of China. The main contents and contributions include: 1) On basis of the defined biorthogonality, we propose several lower-order cost functions for array-signal blind separation. Some efficient fixed-point algorithms for searching the optimal points of the cost function and finding one independent component are established. Multiple independent components can be obtained by using multistage decomposition approaches, even if there is the colored noise; 2) A second-statistics-based beamforming method is proposed; 3) Two method for implementing the blind separation of the nonstationary source signals in the impulse noise environments are studied. 4) We investigate several algorithms for blindly separating the nonstationary source signals; 5) The subspace-based blind identification approaches in the impulse noise environments are researched; 6) A matrix pencil method for second-statistics-based blind equalization are proposed. 18 papers were published, where 9, 7 and 1 papers are or will be cited by SCI, EI and ISTP.

针对实际阵列信号模型(色噪声,非平稳),研究新的循环平稳信号盲分离方法;色噪声抑制及?研究具有一定特征的平稳与非平稳信号的设计,以适应实际的盲信号处理;基于信号的空时特征,利用时频分布和辅助变量的方法,研究在空时色噪声环境下的非平稳信号的盲分离和信道响应的估计,并用于时变信道的均衡;研究基于信号盲分离方法的阵元误差校正.绕过多昀唇Ｎ蟛罴胺植际礁扇藕投嗦肪痘肪扯愿叻直孀允视φ罅行藕糯矸椒ㄊ导视τ孟拗?有重要意义.