基于HMM-BP-DS的雷达装备BIT虚警抑制方法  

False Alarm Reduction Method on Radar Equipment BIT Based on HMM-BP-DS

作　　者：李志强[1] 缪毅[1] 胡文华[1] 黄亮[2] 

机构地区：[1]军械工程学院 [2]中国人民解放军73071部队

出　　处：《测控技术》2015年第6期48-51,共4页Measurement & Control Technology

摘　　要：针对雷达装备BIT虚警"假报"、"错报"的两类表现,提出一种基于HMM-BP-DS的虚警抑制方法。首先将雷达系统状态划分为正常、虚警和故障3个状态,利用HMM区分系统是否产生故障,消除"假报"现象;再对故障状态下采集到的雷达数据采用BP神经网络和D-S证据理论处理,对故障进行定位,以消除"错报"现象。实验分析结果表明,该方法能有效降低雷达装备BIT虚警率。A false alarm reduction method on radar equipment BIT based on HMM-BP-DS is proposed to elimi- nate two appearances of BIT, which are false report and misstatement respectively. Firstly, the states of radar system are divided into three states that are normal, false alarm and fault. Then the hidden Markov model （HMM） is used to diagnose whether the system failures. This step can eliminate the appearance of false report. Secondly, the method applies the BP neural network and D-S evidential theory into the detected radar signal of fault state to locate the faulty part. This step will eliminate the appearance of misstatement. Theoretical analysis shows that this method can effectively reduce false alarm rate of radar equipment BIT.

关 键 词：机内测试 虚警 隐马尔科夫模型 BP神经网络 D-S证据理论 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]