多目标聚类融合跟踪中的特征信息利用

对雷达和电子支援系统的航迹关联问题，利用模糊C-均值聚类算法实现数据关联。该算法综合利用属性信息和动力学信息，通过构造一个影响距离因子函数．根据识别分类结果。调整样本与类中心的距离。最后融合方位角量测得到目标的状态估计。蒙特卡洛仿真表明．新算法比单一使用动力学量测的算法对多目标关联问题更有效。

Clustering Fusion and Characteristic Information Based Multi-target Tracking

For the problem of multi-target tracking using radar and ESM, a fuzzy C-means clustering based data association algorithm was used. A distance influence function was constructed, which made use of characteristic information and dynamical information, and adjusted distance between the measurements and cluster center based on classification. Finally, the state estimation of the targets was gained through measure-to-measure fusion. The Monte Carlo simulation results proved the proposed algorithm is performance better than the algorithm using only dynamical information for multi-target tracking.