摘要
针对联合概率数据关联算法计算量上存在的组合爆炸问题,本文引入最大熵模糊聚类算法实现多目标的数据关联。使用最大熵模糊聚类得到的模糊隶属度表示目标与量测之间的联合互联概率;分析了公共回波对航迹更新的影响,对公共回波的权值进行衰减,对非公共回波的权值进行扩大,避免了航迹合并;此外根据差异因子的特性,给出剔除无效回波的方法,减少了计算量。仿真结果表明,与现有数据关联算法相比,新算法具有更优的跟踪效果。
In the view of high computational load of the joint probabilistic data association algorithm, a new data association algorithm based on maximum entropy fuzzy clustering algorithm is introduced in the paper. As the result of maximum entropy fuzzy clustering, fuzzy membership grade obtained is used as the joint association probability between target and measurement. Effect of public echoes on the update of tracks is analyzed ; track coalescence is prevented by reducing public echoes' weight and increasing non-public echoes' weights. Besides, a method to eliminate invalid echoes is given based on characteristic of discrimination factor; therefore the computational load is reduced. Simulation results show that, comparing with the existing data association algorithm, better tracking performance can be obtained by adopting the new data association algorithm.
出处
《火控雷达技术》
2013年第3期34-38,共5页
Fire Control Radar Technology
关键词
数据关联
最大熵模糊聚类
联合概率数据关联
data association
maximum entropy fuzzy clustering
joint probabilistic data association