Number of blind signal processing CITIC source estimation method
|School||University of Electronic Science and Technology|
|Course||Signal and Information Processing|
|Keywords||Blind Signal Processing Source number estimation Eigenvalue correction Channel Expansion Multistage Wiener filter|
In complex modern electronic communication environment , how to quickly and accurately mix of communications and radar reconnaissance a communications signal processing field one of the problems . Blind source separation techniques can be a good solution to the complex signal environment in the identification and separation of the signal , in recent years in various fields has been widely studied . But to the successful realization of blind signal separation and extraction, you must first mixed signal accurately estimate the number of source signals . Currently more source number estimation applied to array signal processing . In blind signal processing, because of the mixing matrix does not have a particular array flow caused an estimated increase in difficulty . Because blind source separation signal number estimation is the key to an estimated performance directly affects the system performance blind source separation , so the estimated number of blind signal source research on with urgency and significance. The main content of this paper is in the context of blind signal separation mixed-signal communications and radar estimates of the number of questions , the article's core work and innovation is mainly reflected in the following aspects : 1. Summarizes the blind signal processing and source number estimation history and research , detailing several classic source estimation criteria and blind signal mixed environments were analyzed and compared its performance . 2 In the wideband signal model combines multistage Wiener filter technology blind signal processing source number estimation problem , simulation results show that the algorithm for mixed-signal performance of a good estimate of the number of sources . 3 pairs of feature-based decomposition source estimation algorithm in-depth analysis of the correlation matrix for the noise on the eigenvalues ??of the impact studies based on eigenvalue correction techniques to improve the overdetermined mixed environments of the various algorithms in a small SNR the performance . 4 By cumulant array expansion characteristics of the draw, in-depth study based on fourth-order cumulants of the channel expansion technology , simulation of the application channel expansion technology source number estimation algorithm underdetermined mixing conditions the signal source number of detection and correction techniques through a combination of characteristic values ??to improve the algorithm to improve detection algorithm performance. This paper concludes with a summary of the work , and for the next study presents personal opinion .