基于FISST的多源异类信息配准算法  

FISST based dissimilar information of multi-source registration algorithm

作　　者：叶宏[1] 曹学军[2] 李军[1] 李世玲[1] 屈新芬[1] 杨战平[1] 

机构地区：[1]中国工程物理研究院电子工程研究所,四川绵阳621999 [2]中国工程物理研究院应用电子学研究所,四川绵阳621999

出　　处：《太赫兹科学与电子信息学报》2014年第6期865-869,共5页Journal of Terahertz Science and Electronic Information Technology

基　　金：国家自然科学基金NSAF资助项目(U1330133)

摘　　要：针对超高速、超强机动等特定目标的探测和跟踪问题,雷达等传统传感器的量测值少,难于获得有效的探测跟踪,还需要人工情报、专家库信息等异类信息,对于这些信息的配准融合处理尚无有效方法。基于有限集合统计学(FISST)理论,在对几种异类信息进行统一表示并给出相应的滤波处理方法基础上,提出了焦点目标的匹配概率(匹配度)作为异类信息配准指标,并采用卡尔曼证据滤波方法进行计算。最后给出了仿真实例,验证了方法的有效性和可行性。It is hard to detect and track special targets with super high-speed and super strongmaneuverabilityefficiently, only by traditional sensors such as radar. Therefore, intelligence informationand expertise base etc. should be needed. Nevertheless, there are no practicable means for themanagement of multi-source heterogeneous information registration. The characteristics of special targetsare introduced and analyzed firstly. Some kinds of heterogeneous information are expressed in Finite-SetStatistics（FISST） theory and some corresponding filtering algorithms are put forward. The matchingprobability of focused targets is taken as the registration index of heterogeneous information, and theresults are computed with Kalman evidence filter. Numerical simulation validates the feasibility of theproposed method.

关 键 词：多源信息 有限随机集 信息融合 系统配准 

分 类 号：TN953[电子电信—信号与信息处理] V243.2[电子电信—信息与通信工程]