项目名称: 基于多解析度近似和时序分析的传感器网络数据模型研究
项目编号: No.61272410
项目类型: 面上项目
立项/批准年度: 2013
项目学科: 自动化技术、计算机技术
项目作者: 彭凯
作者单位: 华中科技大学
项目金额: 65万元
中文摘要: 传感器网络是物联网的基础,其基础理论与关键技术的研究为当前国际研究热点,涉及信息科学、计算机科学,数学等学科领域的交叉。高效率数据采集和处理是传感器网络研究的关键,传感器数据查询,网络编码、定位覆盖、异常检测等问题给传统的传感器数据模型研究提出巨大的挑战,急需在时空相关性、高维模型参数估计、搜索算法上展开深入研究。本项目针对传感器数据理论模型问题,拟探索构建基于高斯混合模型和隐马尔可夫过程模型的时空相关的多解析度近似数据模型及基于线性网络编码的数据处理模型,由此建立高维数据采样和处理的统一架构模型,并进一步探索采样与处理联合架构模型下的传感器具体应用,形成高效数据处理和采集关键技术,实现在传感器数据理论模型的创新与突破,达到在传感网,物联网产业中的应用。本项目的研究能够为传感器网络信息理论的研究奠定理论基础,积累关键技术,促进传感网,物联网在我国的广泛产业化应用。
中文关键词: 传感器网络;骨架提取;填充曲线;时空相关;多解析度
英文摘要: As the fundamental of the internet of Things, sensor network receives extensive attentions by interdisciplinary experts from information science, computer science, mathermatics and other fields.High efficient data acquisition and processing are the main concerns in sensor network. The significant challenges in data acquisition and processing of sensor network include data inquiry, network coding, positioning, covering and abnormal value detection. It is very desirable to investigate the correlation between time and space, parameter estimation in high-dimensional data model and heuristic searching algorithm to overcome these problems .This project aims to establish a multi-resolution hybrid approximate data sampling model based on a Gaussian mixture function and Hidden Markov Process.It also develops a data processing model by linear network coding method .In this way we propose an unified framework to describe the sampling and processing of high-dimensional data model.We further explore the specific application under the sampling and processing joint model, propose a high efficient data acquisition and processing technology . This research can make a breakthrough and innovation in sensor network data model, which will be practical in sensing network and internet of things.This project can provide basi
英文关键词: sensor network;Skeleton Extraction;Filling curve;time-space correlation;Multi-resolution