A joint data association, registration, and fusion approach for distributed tracking
A joint data association, registration, and fusion approach for distributed tracking
复制标题
DOI：
10.1016/j.ins.2015.06.042
复制
发表时间：
2015-12
期刊：
影响因子：
--
通讯作者：
Hao Zhu;H. Leung;K. Yuen
中科院分区：
文献类型：
--
作者：
Hao Zhu;H. Leung;K. Yuen
本文提出了一种用于分布式跟踪的联合数据关联、配准和融合方法。由于传感器偏置隐含地隐藏在本地轨道后面，因此这里使用伪测量方法来允许在轨道级别进行配准。最大似然函数制定的关联，注册和融合。期望最大化（EM）算法，然后开发执行跟踪注册，关联，融合同时进行。计算机仿真结果表明，该方法具有较好的参数和状态估计性能。
In this paper, a joint data association, registration, and fusion method is proposed for distributed tracking. As sensor biases are implicitly hidden behind the local tracks, a pseudo measurement method is used here to allow registration at the track level. A maximum likelihood function is formulated for association, registration and fusion. An expectation maximization (EM) algorithm is then developed to perform the track registration, association, and fusion simultaneously. Computer simulation results demonstrate the proposed method has an improved parameters and state estimation performance.