多传感器系统估计和判决融合中的维数压缩问题

项目名称： 多传感器系统估计和判决融合中的维数压缩问题

项目编号： No.61473197

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 自动化技术、计算机技术

项目作者： 宋恩彬

作者单位： 四川大学

项目金额： 80万元

中文摘要： 近几十年来，多传感器信息融合在国防、科技、经济、社会等领域发挥着极其重要的作用。为了实现传感器信息融合，各个传感器之间需要通信。在无线传感器网络中，由于通讯总带宽及功率的限制，只能给每一个传感器分配较小的带宽和功率，因此如何预先压缩传感器获得的数据或观测以便节省功耗和通信带宽实现实时处理是多传感器网络的关键问题之一。另外，在信息处理过程中，由于各种条件的限制，人们获得的模型和参数信息大多是不精确的，这种情况下一般采用稳健性方法处理估计和判决问题。本项目研究多传感器系统中稳健估计融合与稳健判决融合的维数压缩问题。本项目将利用国际上判决、估计领域的最新成果，尤其是近年来国际上顶级数学家发展的稳健优化和随机优化的理论和方法，结合所在团队前几轮信息融合基金项目研究中积累下来的方法和经验以及所取得的成果，力争获得多传感器系统中稳健估计融合与稳健判决融合的维数压缩问题的创新性成果。

中文关键词： 信息融合；状态估计；分布式处理；维数压缩；无线传感器网络

英文摘要： In the past few decades, multisensor information fusion has been very important powerful in national defence, science and technology, economy and society.It is necessary that all the sesors communicate with each other for the goal of information fusion. In wierless sensor networks, since the total communication bandwidth and power are limited, the sensors constructing the networks are often equipped with small bandwidth and power, consequently, it is a key issue of multisensor neworks that every each local each sensor compresses its datas or measurements to develop bandwidth-and energy-efficient strategies for various sensor network real-time processing tasks. In addition, for various information procedure, people can only obtain uncertainty information of model and parameter by all kinds of constraints. For the case that the model with uncertainty information, we will deal with it by robust method. Hence, we will researche the dimension reduction topic about robust estimation fusion and robust decision fusion for the multisensor system. In this project, by making full use of the latest progresses of estimation, decision provided by international researchers, in particular, by using robust optimization and stochastic optimization theories and methods developed by some information processing experts and mathematicians, and combining success methods, experience and achievements that we deal with information problem, we will try our best to make great progress for the demension reduction problem of robust estimation fusion and robust decision fusion for the multisensor networks.

英文关键词： Information fusion;State estimation;Distribiuted processing;Dimension reduction;Wireless sensor network