基于混合编码遗传算法的最优跟踪门  

Optimal Tracking Gate Based on Hybrid Encoding Genetic Algorithm

作　　者：赵菡[1] 张琤 林家骏[1] 

机构地区：[1]华东理工大学信息科学与工程学院,上海200237

出　　处：《华东理工大学学报（自然科学版）》2017年第6期844-848,共5页Journal of East China University of Science and Technology

摘　　要：跟踪算法优化可以提高跟踪质量,选择恰当的跟踪门是优化跟踪算法的关键措施之一。本文提出了一种基于混合编码的遗传算法,用于杂波环境下目标跟踪过程中跟踪门参数的离线优化。该算法将二进制编码与浮点数编码结合,对跟踪门的形状和大小进行混合编码,选择跟踪精度性能指标构造遗传算法的适应度函数,以此将跟踪算法的优化问题转化为遗传算法寻优,在不同杂波环境下优化跟踪门参数设置。The track qualities can be depended on tracking algorithm optimization. One of the essential optimization problems in tracking is gate selection. In this paper, a hybrid encoding genetic algorithm is proposed to of--line optimization of tracking gate parameters for maneuvering target tracking in clutter. Binary string and floating-point string represent shapes and size of tracking gate respectively. Hellinger distance is chosen for performance evaluation of tracking and can be core part of the fitness function of genetic algorithm. Generally speaking, the tracking system optimization can be converted into genetic algorithm optimization,then the tracking gate parameters can be efficiently tuned in different scenarios.

关 键 词：目标跟踪 跟踪门 遗传算法 混合编码 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]