顾及目标运动多信息特征的蚁群数据关联方法

A New Ant Colony Data Association method with Multi-feature Considered in Target Moving

针对多目标跟踪的数据关联问题,提出一种将目标运动信息多特征进行融合的蚁群数据关联方法。首先,根据数据关联的具体问题,重新定义了蚁群数据关联方法中路径与路径长度两个概念。其次,考虑目标运动过程中的多种信息特征,即距离信息、方向信息以及灰度信息,将这三种信息特征有机融合,共同作为数据关联标准实现多目标的蚁群数据关联。在多目标跟踪实验中,论文采用EKF滤波方法对目标运动状态进行估计。仿真实验对两个目标交叉运动的情况进行了跟踪估计。实验结果表明,考虑多信息特征后的基于蚁群数据关联的方法在计算量相当的情况下,能较未考虑方向信息以及灰度信息的蚁群数据关联方法获得更高的正确关联率,算法的综合性能优于现有的数据关联方法。

Due to the advantages of ACO（Ant Colony Optimization） in solving complex problems,this paper proposes a new data association algorithm,which is based on Ant Colony Optimization in a cluttered environment,called DDG-ACDA（Distance Direction Gray-ACDA）.In the first instance,the concept for tour and the length of tour are redefined according to the specific requirement for data association problem in multi-target tracking.In the next place,the distance information,directional information and gray information are incorporated into the proposed method,since they are the very important factors that affect the performance of data association process.In the computer simulation of this paper,two targets which move in criss-cross motion are used to validate the performance of the proposed method,and at the same time,the traditional nonlinear filter,EKF,is employed to estimate the target states.Computer simulation results show that the proposed method could carry out data association process in an acceptable CPU time,which is less than other traditional methods,and the correct data association rate is higher than that obtained by the data association algorithm not combined with directional information and gray information.The proposed method is effective for data association in multi-target tracking.