广义似然比检验在APES相干函数估计的应用研究
The Amplitude and Phase Estimation(APES)spectral estimator has been proposed to estimate the magnitude squared coherence(MSC),which will be referred as APES-MSC.The APES-MSC seems to offer better performances than the Welch-based and the Capon-based MSC estimators，but it still suffer from the impact of the noise and the non-correlated signals due to the limited number of available samples.To solve this problem，we give a new insight into the APES-MSC and reveal that the reduced-rank techniques could be applied to the coherence matrix.In this paper，the GLRT-based(generalized likelihood ration test,GLRT) rank detection is chosen to estimate the rank. Numerical simulation results verify that the APES-MSC combined with the GLRT-based rank detection method could effectively improve the estimation accuracy.
非相关信号 相干估计 广义似然比检验 数值模拟
Zheng Chengshi 郑成诗 Xiao Chuansheng 肖椽生 Hu Xiaohu 胡笑浒 Ye Qinghua 叶青华 Li Xiaodong 李晓东
Key Laboratory of Noise and Vibration Research, Institute of Acoustics,Chinese Academy of Sciences, 中国科学院声学研究所噪声与振动重点实验室,北京 100190
国内会议
北京
中文
197-202
2009-10-01（万方平台首次上网日期，不代表论文的发表时间）