基于模糊聚类的多传感器特征关联算法研究

Study on Characteristics Association Algorithm Based on Fuzzy Clustering

特征关联源于量测过程中的不确定性,是无源多传感器多目标跟踪中一个关键环节。模糊集理论的基本思想是把经典集合中的绝对隶属关系模糊化,为处理不确定事物的建模提供有力工具。一种基于模糊聚类的辐射源特征关联模型被提出,同时给出了确定相似性度量和检验门限的方法。最后模拟产生了雷达数据库,对雷达数据库进行了聚类,并与硬聚类算法和灰色聚类算法进行了比较,实验结果证明了该方法的优势和有效性。

Characteristics association stems from the uncertainty in the measurement process,which is the key link of passive multi-sensor and multi-target tracking.The basic idea of fuzzy set theory is fuzzifing the absolute generic relationship of the classical collection and providing a powerful tool to deal with modeling uncertain things.A radiation source characteristics association model based on fuzzy clustering is proposed while a method to determine the similarity measure and the detection threshold is given.Finally,the radar database is simulated and clustered afterwards,hard clustering algorithm and grey clustering algorithm are compared with the fuzzy clustering algorithm,the experimental results demonstrate the advantages and effectiveness of the fuzzy clustering algorithm.