协方差融合算法在时滞系统中的应用研究

项目名称： 协方差融合算法在时滞系统中的应用研究

项目编号： No.61503125

项目类型： 青年科学基金项目

立项/批准年度： 2016

项目学科： 自动化技术、计算机技术

项目作者： 高媛

作者单位： 黑龙江大学

项目金额： 19万元

中文摘要： 时滞系统在信号处理、通信和控制领域广泛存在，时滞系统的状态估计问题一直是近年来研究热点问题。本项目将研究协方差交叉融合算法在时滞系统中的应用，重点研究融合估计问题。对于定常时滞系统，基于现代时间序列分析方法，利用嵌入ARMA模型，将时滞系统转换为带有色噪声的无时滞系统，获得协方差交叉融合估计器；对于时变时滞系统，采取“观测重构+协方差融合”的技术路线，在融合中心获得系统相同时刻状态的融合估计；对于无序量测网络系统，利用序贯协方差融合算法，结合前面的工作基础，获得传感器网络的融合状态估计。并且对于上面提出的各种协方差融合估值器进行精度分析，并用仿真算例进行估值器有效性的验证。通过协方差融合器的获得，为多传感器时滞网络系统的状态估计问题提供一些新的途径、办法。

中文关键词： 协方差交叉融合；分布式；互协方差阵；时滞系统；状态估计

英文摘要： Time-delay system appears extensively in signal processing, comunication and control fields. The state estimation problem of the time-delay system has been a hot spot issue recently. The application on covariance intersection fuison algorithm in time-delay system will be researched in this project, whose focus is the fusion estimation problem. For the time-invariant time-delay system, based on modern time series analysis method, using the Embedded ARMA model, the time-delay system can be changed into the delay-free system with colored noises, and then the covariance intersection fusion estimator can be obtained. For the time-varying time-delay system, the technical route, “measurement reorganization + covariance fusion”, the fused estimation for the state on the same time will be got at the fusion center. For the out-of-sequence measurement network system, by the sequential covariance intersection fusion algorithm, combined with the above working foundation, the fusion state estimation of the sensor network will be gained. Moreover, the precision analysis of these above covariance intersection fusion estimators will be proved, and the validity of these fusion estimators will be verified by some simulation examples. The new ways to the state estiamtion problem for the multisensor time-delay network system will be supplied by our covariance fusers.

英文关键词： Covariance intersection fusion;Distributed;Cross covariance matrix;Time-delay system;State estimation