摘要
提出了一种机动目标的空间群生成算法,该算法的思想是定义空间群目标间的多种特征相似度,通过一级融合获得的信息,计算目标多种特征相似度矩阵并融合求得合群结果。该方法用特征相似度空间属性计算代替了测量空间的目标属性计算,提高了可靠性。该算法的时间复杂度仅为O(n2)。仿真实验表明,该算法具有较高的合群效率,结果可信。
A new group formation algorithm for maneuvering target was proposed. It defined multiple key property similarities by extracting the results of level 1 processing of data fusion and fused to group formation.It realized group formation in real time at multi-levels with lower calculation complexity.Due to the fact that uncertainty of measurement space has been mapped to the fuzzy similarity space,the uncertainty of measurement space was resolved using fuzzy match mechanism.In addition,it was a feasible and reliable approach for space group formation and event detection in battlefield.Simulation results showed that it was efficient to generate group formation for maneuvering target in situation assessment.The time complexity of the algorithm was only O(n~~2).This approach can also be used to solve cluster problem for maneuvering target of other types.
出处
《清华大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2006年第z1期1036-1040,共5页
Journal of Tsinghua University(Science and Technology)
基金
国防"十五"预研基金项目(806030101)
关键词
机动目标
空间合群
态势估计
聚类分析
maneuvering target
space group formation
situation assessment
clustering procedure