多红外传感器融合系统数据关联算法

针对传统多红外传感器多目标多维分配数据关联模型在构造关联代价时未充分考虑位置估计不确定性所引入的随机误差问题，提出了一种可精确到二阶的高精度关联代价构造方法。取非线性量测函数泰勒展开式的前二阶项，将最小二乘估计的均值和方差信息代入得到伪量测信息的均值和方差，继而将伪量测与真实量测信息的统计距离作为最终的关联代价。最后对不同关联算法的正确率进行了实验对比，仿真结果表明修正后的关联代价能够更精准地反映出数据关联的可能性，基于该修正代价的关联算法较之其他关联算法可获得更好的关联性能。

Data association algorithm for multi-infrared-sensor system

When established association cost, traditional multi-dimension assignment data association algorithm for multi-infrared-sensor ignored the random errors introduced by least square estimation. To overcome the problem, a modified cost function that can get second-order accuracy was proposed. The first two items of Taylor series for nonlinear measurement function was kept down. The pseudo measurements can be got by using the mean and covariance of position estimation according to the preserved series. Then the statistical distances between real measurements and pseudo ones worked as the association costs. Finally, the correct data association ratio of the several association algorithms were compared through simulation experiments. Simulation results show that the modified cost function can reflect the association probability more accurately and the algorithm based on it can achieve better performance than the others.