改进并行集中式数据关联算法

针对目标批次过多导致计算上的组合爆炸问题，提出一种改进并行集中式多传感器不敏近似联合概率数据关联（joint probabilistic data association，JPDA）算法。该算法首先采用基于unscented变换的卡尔曼滤波（unscented Kalman filter，UKF）实现非线性系统中状态估计的递推,然后通过改进的并行集中式方法将数据传至中心，利用改进的JPDA方法进行量测点迹与目标航迹关联。仿真实验表明，该算法在非线性复杂环境中具有较好的数据关联正确率，且计算耗时较集中式串行不敏JPDA算法少。

A data association algorithm of improved parallel centralized

In order to solve the problem of the N-P hard with the increased number of targets, a novel unscented joint probabilistic data association based on centralized multisensor algorithm is proposed. First, UKF is used for the propagation of state estimate in the nonlinear system.Then the data is transmitted to the fusion center through the use of the reformative parallel centrlizad way. Finally the improvement JPDA is used to associate the data. Simulation results show the algorithm has a higher accurate rate of association in a clutter environment, and the CPU occupies a relatively short time.