一类模式观测时滞Markov切换系统融合.pdfVIP

th Proceedings of the 25 Chinese Control Conference 7-11 August, 2006, Harbin, Heilongjiang 一类模式观测时滞Markov切换系统融合 魏莉莉, 潘泉, 梁彦, 程咏梅 西北工业大学 自动化学院，陕西西安 710072 E-mail: quanpan@nwpu.edu.cn 摘 要: 随着现代传感器处理器的发展及对系统性能要求的提高，本文提出模式观测滞后状态量测下的Markov 切换系统融合问题。并针对模式观测滞后状态量测一步时的融合问题，通过状态扩维的方法，转化成无时滞 Markov切换系统的广义滤波问题。基于贝叶斯估计理论，推导滤波公式，给出了模式观测增强的交互式多模型 广义滤波算法。机动目标跟踪的仿真算例表明其有效性。 关键词: 目标跟踪，时滞，Markov切换系统，信息融合 Fusion for a class of Markov switching system with delayed mode measurements Wei Lili, Pan Quan, Liang Yan, Cheng Yongmei Northwestern Polytechnical University, College of Automation, Shaanxi,710072,China E-mail: E-mail: quanpan@nwpu.edu.cn Abstract: With the development of sensors and requirement for better system performance, an information fusion problem for a class of Markov switching system with delayed mode measurements is put forward. What’s more, when mode measurement is one-step delayed relative to state measurement, this kind of fusion problem is changed into a generalized filtering problem for a common Markov switching system through state-augmentation approach. This algorithm (mode-measurement-enhanced interacting multiple model generalized filtering algorithm, namely mme-IMM-gfa) is derived by Bayes theory, and is illustrated via a maneuvering target tracking simulation example. Key Words: target tracking, time-delay, Markov switching system, information fusion 并假设。在这众多算法中，IMM应用最为广泛。以 1 引言(Introduction) IMM为基础，形成了比较完善的平滑算法和一些新 型的附带模式观测通道的滤波算法。 混合系统[1]中