文献类型：会议论文

英文题名：An improvement to variable-structure multiple-model algorithm for maneuvering target tracking

作者：Zhang, Wenjie Long, Fei Liu, Xia Ding, Adan

第一作者：Zhang, Wenjie

机构：[1] College of Big Data and Information Technology, Guizhou University, Guiyang, China; [2] College of Information Engineering, Guizhou Institute of Technology, Guiyang, China

第一机构：College of Big Data and Information Technology, Guizhou University, Guiyang, China

会议论文集：Innovation in Design, Communication and Engineering - Proceedings of the 3rd International Conference on Innovation, Communication and Engineering, ICICE 2014

会议日期：October 17, 2014 - October 22, 2014

会议地点：Guiyang, Guizhou, China

语种：英文

外文关键词：Clutter (information theory) - Kalman filters - Probability

年份：2015

摘要：The Variable-Structure Multiple-Model (VSMM) estimation is a popular method that is used widely in maneuvering target tracking. In this paper, based on the combination framework of Expected-Model Augmentation (EMA) algorithm and Minimal Model-Group Switching (MMGS) algorithm, two improvements will be employed for tracking strong maneuvering target. Because the modelgroup switching algorithm can not be activated and terminated the models without in the total model-set designed previously, model set adaptation is limited more or less. However, EMA algorithm can remedy this drawback. Thus, the combination of two algorithms will increase the estimation performance obviously. Besides, standard Kalman filter lacks adaptive ability and the transition probabilities among model set are usually invariable. So, strong tracking filter and a nonhomogeneous transition probability matrix are utilized by replacing the standard Kalman filter and transition probability matrix designed in a priori. Several simulation results used to evaluate the performance of the proposed algorithm show the effectiveness of it compared with other three algorithms. ? 2015 Taylor & Francis Group, London.