摘要
在研究联合概率数据互联(JPDA)法的基础上,探讨了一种近似的JPDA方法,该方法引入"近似聚"的概念,用近似聚矩阵代替确认矩阵,把一个比较大的确认矩阵分成若干个互不相交的小的确认矩阵来进行互联矩阵的拆分,以此来减小确认矩阵的观测量数和回波数,从而降低了算法的复杂度和计算量,使JPDA算法适应密集目标实时跟踪的需要.此外,还分析了并行和串行2种多传感器联合概率数据互联法,并就算法的复杂度进行了分析比较.
To meet the requirement of tracing large quantities of targets real time, an approximate multison-sor joint probabilistic data association(JPDA) algorithm were discussed after studying the joint probabilistic data association (JPDA) methods. This method adopts the conception of approximate cluster which uses approximate ctuster matrix to get the association matrix and breaks the association matrix by dividing a large association ma- trix into some small association matrixes of non-intersection each other. This method reduces the counts of observed and echo and the complexity and quantities of the algorithm. In addition, two kinds of algorithms, parallel and serial, of muhisensor joint probabilistic association algorithms were analyzed and compared.
出处
《光电技术应用》
2009年第4期25-28,53,共5页
Electro-Optic Technology Application
关键词
数据融合
联合概率
多传感器
data fusion
joint probabilistie association
multi-sensor