|
|
|
| Parametric Multichannel Target Detection in Heterogeneous Environment |
| Shang Xiu-qin①② Song Hong-jun① Xu Hai-sheng①② Zheng Jing-bo①② |
| ①(Institute of Electronics, Chinese Academy of Sciences, Beijing 100190, China) ②(Graduate University of the Chinese Academy of Sciences, Beijing 100049, China) |
|
|
|
|
Abstract Parametric multichannel target detection in heterogeneous environment is studied in this paper, where the disturbances are represented by Vector AutoregRessive (VAR) model with its spatial covariance matrix following complex inverse Wishart distribution with degrees of freedom μ and mean Q. When Q is unknown, the Neyman-Pearson Parametric Adaptive Matched Filter (NP-PAMF), Bayesian PAMF (B-PAMF) and its normalized version (B-NPAMF) are proposed based on NP detection rule. And when it is known, the maximum a-posteriori PAMF (MAP-PAMF) and its normalized version (MAP-NPAMF) are proposed followed MAP decision rule. It is shown that NP-PAMF and B-PAMF are both dependent onand B-PAMF is convergent to the PAMF when μ→∞; B-NPAMF has no relation with μ and is consistent with the classic NPAMF. In MAP-PAMF, the MAP estimator of the spatial covariance matrix consists of the classic estimator and the prior knowledge, and the weigh of the later is controlled by μ. Finally, the complex issues and the detection performances are nalyzed, showing that: Bayesian parametric detectors possess good performances and they are better than their normalized counterparts.
|
|
Received: 03 September 2010
|
|
|
|
Corresponding Authors:
Shang Xiu-qin
E-mail: shangxiuqin2009@gmail.com
|
|
|
|
|
|
|
2011, Vol. 33 
