摘要
针对传统目标分群单一算法中存在的适用范围小、战术意义不明显等问题,将自底向上逐层深入的目标群形成过程进行重新定义,提出了一种新的态势分析中目标群特征识别的处理方法。首先,将通过深层次态势分析得到的目标执行的战术任务状态属性作为首要分群依据,并采用自组织神经网络对目标进行空间任务群识别;其次,模型以兵力间信息协同增益效应对相互作用与战术关联群形成的影响为基础,计算空间任务群的多种属性相似度矩阵,并融合求得合群结果。仿真实验对比结果表明,该方法可有效减小空间上的错误分群率,并识别出具有战术联系的空间群。
Due to the limited range and the tactical significance obscurely of the traditional target group recognition, it redefines targets grouping process hierarchically and presents a novel method to deal with the recognition of target group characteristic in situation assessment. Firstly, it takes the tactical tasks of operation platform as the chief grouping evidence, which gets from situation analysis deeply, and adopts Self-Organizing feature map （SOFM） neural network to recognize the space task group. Then, Based on information coordinative plus exercises influence on interactional group and tactical associate group, it computes the attribute comparability matrix of space task group to get the incorporate groups. The result indicates that the method can reduce the error rate of clustering.
出处
《指挥控制与仿真》
2015年第4期25-30,共6页
Command Control & Simulation
关键词
态势分析
目标群
群特征识别
多相似性测度
situation analysis
targets group
recognition of target group characteristic
multi-comparability measure
