针对传统数据关联算法存在计算量偏大或关联精度不高的问题,提出了一种利用遗传模糊聚类策略来求解数据关联问题的算法。该算法将多传感器多目标的数据关联问题看作是一类约束条件下的组合优化问题,先通过对同一时刻不同传感器提供的量测按照其相似性用遗传算法进行模糊聚类,再用聚类后的等效量测对各目标的状态进行估计。聚类方式的改进不仅增加了算法的局部寻优能力,有效地减少了计算的复杂度,而且还具备一定的野值剔除能力。仿真结果表明该算法关联精度较高,计算量适中,具有一定的工程应用价值。

Traditional data association algorithm has the drawbacks of large computation cost and low association precision. To solve the problem,we proposed a genetic fuzzy clustering based data association algorithm. Taking the multi-target,multi-sensor data association as a combination optimization problem under a type of restrictions,fuzzy clustering were made for the measurement data of different sensors at the synchronized instant according to their similarity,and then the clustered equivalent measurement was used to estimate the state of each target. The improvement of clustering can strengthen the algorithms local optimization capability and reduce the computation complexity considerably while eliminating some outliers. The simulation results show that the presented algorithm has high association precision and acceptable computational complexity,thus is practically valuable.