摘要
分析了舰船噪声阵列声纳观测信号的多途卷积混叠特性和盲源信号分离问题及扩展信号瞬态混合形式下的四阶累积量联合对角化方法。研究了适合多源多途水声信号盲分离技术及波形恢复、修改后的算法对抑制环境噪声的有效性。用仿真信号和水池实验实录信号对修改后的盲分离算法进行了检验,取得了满意的效果。
In this paper, multipath convolutive mixtures of the sonar array observation signals of ship noise and blind separation problems are analyzed. An updated approach is developed to research blind separation and waveform recovery for multiple-source and multipath underwater acoustic signals by using extended joint approximate diagonalization of eigen-matrices (JADE) based on the 4th-order cumulants. The modified algorithm is very effective for reducing the environment noise. Simulation results are given, illustrating a good performance by using both synthesized signals and recorded signals in the water-tank experiment.
出处
《系统工程与电子技术》
EI
CSCD
北大核心
2003年第9期1058-1060,1083,共4页
Systems Engineering and Electronics
基金
国家自然科学基金资助课题(60072049
60172072
30170259)
关键词
高阶累积量
盲分离
舰船噪声
水声信号
High-order cumulant
Blind separation
Ship noise
Underwater acoustic signal
