SIA OpenIR  > 工业信息学研究室
RFID智能信息处理关键问题研究
Alternative TitleResearch of Key Problems on RFID Intelligent Information Processing
库涛1,2
Department工业信息学研究室
Thesis Advisor朱云龙 ; 胡琨元
ClassificationTN911.23
KeywordRfid 流数据处理 复杂事件处理 智能信息处理 面向服务体系结构
Call NumberTN911.23/K58/2009
Pages134页
Degree Discipline机械电子工程
Degree Name博士
2009-01-15
Degree Grantor中国科学院沈阳自动化研究所
Place of Conferral沈阳
Abstract无线射频识别(Radio Frequency Identification,RFID)技术是通过无线射频方式进行非接触双向通信的自动化识别与数据获取技术,在多目标识别、移动目标识别和单品识别等应用领域具有极大的优势,广泛应用于国防、制造、医疗、零售、物流等诸多领域,被誉为21世纪最热门的技术之一。近年来RFID技术受到工业界和学术界高度关注,理论研究和实践应用得到了突飞猛进的发展,同时,在面向RFID海量数据的智能信息处理方面也带来了巨大的挑战。因此,本文以面向应用的RFID数据处理为背景,对RFID智能信息处理及其关键问题进行了深入研究,主要研究内容包括以下几个方面:首先,回顾并总结了流数据处理、复杂事件处理和信息融合等理论方法在RFID信息处理上的主要研究成果和典型方法,为RFID智能信息处理体系的建立奠定了理论依据,针对当前RFID信息处理方法的不足,从融合多领域相关技术,进行综合性、大规模的RFID信息处理的角度,提出了综合方法集成框架,并给出了框架相关模块的构成说明,为进一步提出RFID智能信息处理方面新的理论和方法提供了研究思路。 其次,以构件技术和中间件技术为背景,详细分析了现存的RFID中间件系统的优势与不足,在此基础上,主要针对RFID海量数据处理、RFID多协议数据转换、复杂事件处理以及面向RFID的企业信息集成等问题,提出了RFID智能信息处理软件体系结构,设计并实现了流数据处理引擎、多协议处理器、监控管理器、事件管理器和复杂事件处理引擎,并给出了相关模块的处理算法,从软件体系结构上保证了RFID信息处理软件系统的智能性和可扩展性,软件实现验证了所提软件体系的有效性。再次,在总结和分析流数据管理系统和流数据查询语言方面主要研究成果的基础上,从重用和集成现有流数据处理方法,分离流数据源和处理逻辑,简化软件开发与部署,实时监控与优化RFID流数据处理的研究角度,提出了面向服务体系的RFID流数据处理服务体系,给出了RFID流数据处理服务模型和服务逻辑模型,并给出了基于XML的服务单元描述,数据流描述和服务单元激活描述。最后给出了软件实现,验证了所提出的流数据处理服务的有效性。最后,以事件驱动系统和主动数据库系统为背景,介绍了复杂事件处理所涉及的关键技术,并在分析目前研究的优势与不足基础上,针对面向RFID海量事件的语义信息处理需求,提出了RFID复杂事件模型和复杂事件处理网络模型,并提出了基于时空关系的操作算子和复杂事件模式定义。在分析和总结已有复杂事件探测体系不足的基础上,进一步设计了RFID复杂事件探测中心结构,并给出相关构件的功能描述。对于复杂事件探测问题,主要提出了基于分发策略和消耗策略的探测优化,并给出了策略选择算法,在以上算法结果的基础上提出了基于策略选择的复杂事件探测算法,同传统复杂事件探测算法比较分析显示所提算法的有效性。
Other AbstractRFID (Radio Frequency Identification, RFID) technology as one of the hottest technologies at the 21st century is an automatic identification and data capture technology, which has great advantages in the applications of multi-target recognition, moving target identification, and item identification. RFID technology highly concerned by the industry and academia in recent years has been widely used in national defense, manufacturing, health care, retail, logistics and many other fields, which has been rapidly developed on the theoretical study and practical application. At the same time, intelligence information processing for RFID massive data has also brought great challenges. Therefore, the thesis studies depthly the key problems on the RFID Intelligent Information Processing (RIIP) taking RFID data processing of application-oriented as background. The main contents include the following: Firstly, presents a detailed analysis for current technology, methods, typical software systems, and algorithms of RFID information processing from three aspects of stream processing, complex event processing, and knowledge acquisition, which lay a theoretical foundation for RFID Intelligent Information Processing (RIIP). Based on the above analyses, we point out the lacks of the current research and put forward the new research roadmap called integration framework of synthetic methods for RIIP from the point of integrating RFID data streams,event streams, and knowledge streams. Secondly, proposes the software architecture of RFID intelligent information processing according to the problems of RFID massive data, multi-protocols conversion, as well as knowledge acquisition based on complex event processing. We use component technology to design and implement the RFID stream data-processing engine, multi-protocols processor, monitor manager, event processor, and complex event processing engine, and evaluate performance of related component algorithms. The proposed architecture gives the completed process of intelligent processing from the original RFID data acquisition, protocols conversion to knowledge acquisition, and ensures intelligence, real-time, scalability, and flexibility on the RFID information processing from views of software architecture. Thirdly, proposes a service-oriented RFID stream processing architecture based on reusing the existing stream processing approachs, separating processing logics from stream sources, simplify software development and deployment, and monitoring and optimizing on real time. Addressing the problems of continuous processing on RFID massive data, we propose models and optimization strategies of RFID stream data processin service. Implemation and related performance analysis show that proposed architecture have the high availability and reliability. Finally, proposes RFID complex event model based on the summary technologies on the complex event processing, event-driven and initiative database, and presents the definitions of complex events, operators, and patterns based on relationship of temporal-spatial. Further, proposes model of RFID complex event network from view of distributed events processing, and gives the formal difinitions. According to advantages and disadvantages on existing complex event detection architecture, we propose complex event detection center, and analyze and summarize complex events detection strategies. The algorithm for strategies selection based on heuristic is proposed, and simulation analyses verify validity of the proposed method.
Language中文
Contribution Rank1
Document Type学位论文
Identifierhttp://ir.sia.cn/handle/173321/410
Collection工业信息学研究室
Affiliation1.中国科学院沈阳自动化研究所
2.中国科学院研究生院
Recommended Citation
GB/T 7714
库涛. RFID智能信息处理关键问题研究[D]. 沈阳. 中国科学院沈阳自动化研究所,2009.
Files in This Item:
File Name/Size DocType Version Access License
10001_20051801471902(8797KB)学位论文 开放获取CC BY-NC-SAApplication Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[库涛]'s Articles
Baidu academic
Similar articles in Baidu academic
[库涛]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[库涛]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.