复合高斯杂波中知识辅助检测器的先验信息感知方法

Prior information cognitive method for knowledge aided detector in compound Gaussian clutter

利用先验信息可以提高雷达目标的探测能力,若先验信息与当前探测环境不匹配,知识辅助检测器的性能会受到影响.本文针对Bayes框架下的复合高斯杂波中的知识辅助检测器,提出了杂波纹理分量先验分布参数的感知方法.首先阐述了知识辅助检测器先验信息感知的一般方法.然后针对基于杂波纹理分量先验信息的知识辅助检测器结构,建立了先验模型参数失配与知识辅助检测器检测性能之间的量化关系.进一步利用知识辅助检测器对当前杂波场景进行探测,获得检验统计量和虚警率测量值,从而构造纹理分量分布参数的约束关系.通过分析多个约束关系的交点,获得杂波纹理分量先验分布参数的感知值.计算机仿真分析了这种感知方法的可行性,并利用实测杂波数据对感知方法进行了验证.通过知识辅助检测器检测性能对比分析,采用感知方法获得的先验信息模型参数能够进一步提高检测器的性能.

The radar targets detection performance can be improved by using some prior information, but the mismatches between the prior information and current detection environments may result to the degradation of the knowledge aided detector performance. In this issue, a cognitive method of the clutter texture distribution parameters is proposed for the knowledge detector in compound Gaussian clutter under the Bayesian framework. Firstly, the general cognitive method of prior information for the knowledge aided detection is explained. Then for the knowledge aided detector using prior information of clutter texture component, the quantized relationship between the mismatches of prior information and detector＇s performance is given. Furthermore, using the test statistics and measurement of false alarm probability obtained by the knowledge aided detector under current clutter environment, the constrained relationship of texture component distribution parameters is constructed. The cognitive value of these clutter texture distribution parameters can be calculated by the intersection analysis of nmltiple constrained relationships. The computer simulation is used to analyze the feasibility of cognition method, and the real clutter data sets are used to validate the cognitive method. The detection performance of knowledge detector using cognitive value can be improved furthermore.