基于UMVE算法的分布式模糊恒虚警系统  

Distributed Fuzzy CFAR System Based on Unbiased Minimum-variance Estimation Algorithm

作　　者：郝程鹏[1] 刘斌[1,2] 理华[1,2] 

机构地区：[1]中国科学院声学研究所,北京100190 [2]中国科学院研究生院,北京100039

出　　处：《系统仿真学报》2009年第13期3993-3996,共4页Journal of System Simulation

基　　金：国家自然科学基金(60802072);国防十一五预研基金资助项目(1010602010102)

摘　　要：基于模糊逻辑和无偏最小方差估计(UMVE)算法提出一种新的分布式CFAR系统(分布式模糊UMVE),系统中每个UMVE-CFAR检测器根据参考单元计算出映射到虚警空间的隶属函数值,并传给融合中心,融合中心根据这些值采用模糊融合准则进行判决。仿真结果表明,分布式模糊UMVE在均匀背景和非均匀背景下具有比分布式模糊OS更好的检测性能,同时,模糊算术积融合比二元"与"融合和"或"融合提供更好的检测性能。该研究成果可应用于警戒雷达系统。A new type of distributed CFAR scheme （distributed fuzzy UMVE） based on fuzzy logic and Unbiased Minimum- Variance Estimation （UMVE） algorithm was proposed. In this scheme, each UMVE-CFAR detector computed the membership function value mapping to the false alarm space from the samples of reference cells, and transmitted it to the fusion center. These values were combined to make a decision according to fuzzy fusion rules. Simulation results show that the detection performance of the distributed fuzzy UMVE-CFAR is better than that of the distributed fuzzy OS-CFAR both in homogenous environments and in nonhomogenous environments, and fuzzy algebraic product fusion provides better detection performance than binary ＇or＇ fusion and binary ＇and＇ fusion. This scheme can be used in earlywaming radar system.

关 键 词：检测 分布式恒虚警 无偏最小方差估计 模糊融合 隶属函数 

分 类 号：TN951[电子电信—信号与信息处理] TN957.51[电子电信—信息与通信工程]