航空学报 > 2022, Vol. 43 Issue (5): 325236-325236   doi: 10.7527/S1000-6893.2021.25236

基于改进的FCM和信息熵修正的航迹融合算法

甄绪, 刘方   

  1. 国防科技大学 自动目标识别重点实验室, 长沙 410005
  • 收稿日期:2021-01-08 修回日期:2021-02-06 发布日期:2021-02-24
  • 通讯作者: 刘方 E-mail:smartlf@sina.com

Track fusion algorithm based on improved FCM and information entropy correction

ZHEN Xu, LIU Fang   

  1. National Key Laboratory of Science and Technology on Automatic Target Recognition, National Defense Science and Technology University, Changsha 410005, China
  • Received:2021-01-08 Revised:2021-02-06 Published:2021-02-24

摘要: 在局部航迹信息质量不均衡条件下,选择所有局部航迹进行航迹融合的算法会造成系统航迹质量下降。为了提高跟踪性能,提出了一种基于改进的模糊C均值(FCM)和信息熵修正的航迹融合算法。通过交互式多模型(IMM)滤波后的航迹信息对聚类数据做"质量"修正,改进后的FCM算法对局部航迹进行聚类分析,利用信息熵和隶属度对局部航迹进行选择和融合,达到修正聚类中心和提高系统航迹质量的效果。仿真结果表明:当多个传感器跟踪机动目标时,在传感器的观测精度发生变化和存在量测丢失的情况下,该算法的跟踪性能优于已知的航迹融合算法。

关键词: 不完备信息, 信息熵, 隶属度, 航迹选择, 航迹融合, 模糊聚类

Abstract: When the quality of local track information is not balanced, the algorithm of choosing all local tracks for track fusion will lead to degradation of track quality. To improve tracking performance, a track fusion algorithm is proposed based on the improved Fuzzy C-Means (FCM) and information entropy correction. The "quality" of the clustering data is modified by the track information filtered with the Interactive Multi-mModel (IMM). The improved FCM algorithm is used to cluster the local track. The local tracks are selected and fused by using information entropy and membership degree to modify the clustering center and improve the quality of the system track. Simulation results show that when multiple sensors are used to track maneuvering targets, the tracking performance of the proposed algorithm is better than the known track fusion algorithms in the case of changing of sensor observation accuracy and measurement loss.

Key words: incomplete information, information entropy, membership, track selection, track fusion, fuzzy clustering

中图分类号: