改进的DBSCAN算法在室内多扩展目标跟踪中的研究  

Research on improved DBSCAN algorithm in indoor multiple extended target tracking

作　　者：王国林 于映[1] WANG Guolin;YU Ying(College of Integrated Circuit Science and Engineering,Nanjing University of Posts and Telecommunications,Nanjing 210023,China)

机构地区：[1]南京邮电大学集成电路科学与工程学院,江苏南京210023

出　　处：《电子设计工程》2024年第8期139-143,共5页Electronic Design Engineering

摘　　要：基于密度聚类(DBSCAN)的算法由于不需要提前确定聚类的簇数目且具有抗噪声等特点,被广泛用于雷达多目标跟踪中。针对雷达获取室内目标的量测数据量增大、杂波分布密集导致目标跟踪系统复杂度增加、耗时严重的情况,提出一种改进的方法。该方法通过构建KD树来加速DBSCAN算法中邻域点的查找过程,极大地提高了运算效率,改善了目标跟踪系统的实时性。并且运用Murty算法和无迹卡尔曼滤波算法(UKF)来提高传统多目标跟踪系统的跟踪效率和精度。通过仿真实验将传统DBSCAN算法与K-means算法以及改进的算法相比较,实验结果表明在保证室内多目标跟踪系统的跟踪精度的同时,改进算法极大地提高了跟踪系统的实时性。Density-based spatial clustering of applications with noise(DBSCAN) algorithm are widely used in radar-multi-target tracking because there is no need to determine the number of clusters in advance,anti-noise and other characteristics.In order to solve increased data measurement data for radar obtaining indoor targets and densely dense mixed wave distribution resulting in increasing the complexity of the target tracking system and serious consumption,this article proposes a improvement method to use the construction of the KD tree to accelerate the DBSCAN algorithm point of the domain point of the field of the DBSCAN algorithm.The search process has greatly improved the operational efficiency and the realtime nature of the target tracking system.And use the Murty algorithm and UKF algorithm to improve the tracking efficiency and accuracy of the traditional multi-target tracking system.The traditional DBSCAN algorithm is compared with the K-means algorithm and the improved algorithm through simulation experiments.The experimental results show that while ensuring the tracking accuracy of the multi-target tracking system in the room,the improved algorithm greatly improves the real-time nature of the tracking system.

关 键 词：DBSCAN KD树 目标跟踪 UKF 

分 类 号：TN953[电子电信—信号与信息处理]