SIA OpenIR  > 工业信息学研究室
RFID系统优化模型与智能算法研究
Alternative TitleResearch on RFID Systems Optimization Models and Intelligent Algorithms
陈瀚宁1,2
Department工业信息学研究室
Thesis Advisor朱云龙 ; 胡琨元
ClassificationTN911.23
Keyword射频识别技术 Rfid网络 智能计算 群体智能 模糊聚类
Call NumberTN911.23/C45/2009
Pages144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射频识别技术(Radio Frequency Identification, RFID)作为采集与处理信息的高新技术和信息化标准的基础,被列为本世纪十大重要技术之一。但是,RFID技术的大规模实际应用仍处于探索阶段,RFID系统的应用基础技术还存在着大量尚未解决的关键问题,其中RFID系统优化是RFID技术研究和应用的重要课题。由于RFID系统本身的动态性和不确定性, RFID系统优化面对的一般是非线性、多目标、大规模的复杂优化问题,传统的数学优化算法在处理这些问题时,存在困难。为此,研究新的优化算法成为RFID技术实际应用和理论研究中必须解决的课题。智能计算方法是求解复杂RFID系统优化问题的一种可供选择的算法。智能计算作为一个新兴领域,其发展已引起了多个学科领域研究人员的关注,目前已经成为人工智能、经济、社会、生物等交叉学科的研究热点和前沿领域。智能计算的各类算法已在传统NP问题求解及诸多实际应用领域中展现出其优异的性能和巨大的发展潜力。本文旨在对RFID系统的各种优化问题进行深入研究和探讨,面向RFID技术的实际应用需求构建其优化模型,并基于智能计算思想设计能够有效求解这些复杂模型的新型智能优化算法。具体研究内容包括:首先,进行了RFID读写器网络的调度问题研究。在深入分析RFID网络中读写器冲突类型和成因的基础上,考虑RFID网络中的读写器冲突约束,以最小化系统中的频道数量、时隙分配以及总处理时间建立了RFID读写器网络调度的数学优化模型。从生物学的角度出发提出基于生态捕食模型的改进PSO算法(Particle Swarm Optimizer based on Predator-prey Coevolution, PSOPC),在一定程度上解决了PSO算法在迭代后期随着多样性丧失而陷入局部最优的缺点。应用PSOPC设计了求解RFID读写器网络调度模型的智能求解算法,分别给出算法的求解框架、关键步骤的实现机制。通过在不同规模的RFID读写器网络上进行实例仿真,验证了算法的有效性和模型的正确性。其次,进行了基于菌群自适应觅食算法RFID网络规划问题的研究。考虑RFID系统在不同应用环境下的系统需求,建立了RFID网络规化的数学模型,其目标函数分别为:RFID网络标签覆盖率的最大化目标函数、RFID读写器冲突的最小化目标函数、RFID网络运行的经济效益最大化目标函数、RFID网络运行的负载平衡目标函数以及同时考虑全局目标的混合目标函数。将自然界生物觅食所采用的自适应搜索策略与细菌的趋化行为和群体感应机制相集成,提出了适合求解复杂RFID网络规划问题的菌群自适应觅食算法(Adaptive Bacterial Foraging Optimization, ABFO)。通过仿真实验基于ABFO算法分别对RFID网络规划模型中的五个目标函数进行了实例求解和分析,测试结果与标准PSO算法和遗传算法进行了比较分析。再次,进行了基于系统智能方法的RFID网络规划分布式决策模型研究。采用分布式决策的思想建立了RFID网络规划的层次模型,在一定程度上缓解、分散了RFID网络规划问题的复杂性,以解决具有混合变量(包括离散变量和连续变量)的多目标RFID网络规划问题。针对层次模型求解的复杂性,以复杂适应系统理论为指导思想设计了一种新型系统智能优化算法对RFID网络规划的层次模型进行求解。系统智能算法将群体智能中的单层群体系统概念扩展为多层涌现系统,仿真实验表明新提出的算法显著提高了智能计算方法的寻优能力,以及算法的适应性、鲁棒性和平衡性等性能。最后,进行了RFID网络目标跟踪系统中的数据融合研究。以基于RFID技术的目标定位与跟踪系统为应用背景,提出了基于模糊聚类方法的多RFID读写器数据融合模型框架。通过深入分析蜜蜂采蜜的基本生物学规律,对蜜蜂的个体行为及群体行为进行模拟,提出了一类新型群体智能优化算法-蜂群优化算法(Bee Swarm Optimization, BSO),并将BSO算法嵌入RFID目标定位跟踪系统,作为其模糊聚类的基本算法。仿真研究表明,提出的融合模型能够有效的过滤读写器对跟踪目标的错误监测数据,显著提高目标定位与跟踪的精度。
Other AbstractAs the basis of the informationization standards and the high-and-new technology of information acquisition and processing, RFID technology is listed as one of ten most important technologies of this century. However, the large-scale practical application of the RFID technology is still at the initial stage, and there are many key questions need to be solved for the foundation technology of RFID. Specially, the RFID System Optimization is an important topic of RFID technology research and application. Due to the dynamic and uncertain characteristics of RFID system, the RFID System Optimisation generally faces the misalignment, multi-objectives, large-scale complex optimized questions, which are difficult for the traditional mathematics optimization algorithm to deals with. Thus, the development of the new high-performance optimization algorithms has become the essential issue of the practical application and the fundamental research for RFID technology. The Intelligence Computational method is the selective algorithm for solving the complex RFID System Optimization problems. As an emerging field, the development of Intelligent Computation has aroused many researchers’ interests in many fields. The Intelligent Computation has already become the research focus and the front field of the interdisciplinary studies such as Artificial Intelligence, the Economics, the Sociology, and the Biology interdisciplinary studies. Each kind of algorithm of the Intelligent Computation has already demonstrated excellent performance and the huge development potential in many practical application fields such as the traditional solution of NP-hard problems. This paper gives deep research and discussion about the RFID System Optimization problems, constructs the optimization models according to the RFID practical application demands, and designs the new intelligent optimization algorithms to solve these complex models effectively based on the Intelligent Computation methodology. The concrete research content includes: First, the RFID reader network scheduling problem is studied. On the basis of analysis of the conflict types and causes of the RFID reader networks, this paper establishes the scheduling optimization model of RFID reader networks that considering the reader-writer conflict restraint of the RFID networks. This scheduling model is in order to minimize the channel quantity, the time slot assignment, and total processing time of the RFID system. From the perspective of biology, an improved PSO algorithm is proposed - the Particle Swarm Optimizer based on Predator-prey Coevolution (PSOPC). This algorithm is designed to get over the shortcoming that the PSO algorithm falls into local optima with loss of the diversity in the later optimization period. The solution algorithm of RFID reader networks scheduling models is designed based on PSOPC algorithm. The simulation studies, which compared to Canonical PSO algorithm, show that the PSOPC obtains superior RFID reader scheduling solutions than PSO methods in terms of optimization accuracy and convergence speed. Second, the RFID Networks Planning based on Adaptive Bacterial Foraging Optimization (ABFO) is studied. Considering the system demands in different application environment of RFID system, the RFID Networks planning mathematical model is established. And its objective functions are the maximizing the cover rate of RFID tags, minimizing the RFID reader conflict, maximizing the economic benefit of RFID system, load balancing of RFID readers, and the combined measurement, respectively. In order to improve the BFO’s performance on complex optimization problems with high dimensionality, we apply two natural foraging strategies, namely the Producer-Scrounger Foraging and the Area Concentrated Search, to the original BFO, resulting in two new Adaptive Bacterial Foraging Optimization (ABFO) models, namely ABFO1 and ABFO2. Instead the simple description of chemotactic behavior in BFO, the proposed algorithms can also adaptively strike a balance between the exploration and the exploitation of the search space during the bacteria evolution, by which the significant improvement can be gained. Through the simulation experiment based on the ABFO algorithm, the example solution and analysis about the five objective functions of RFID network planning models are given, and the comparative analysis between the test result and standard PSO algorithm, the Genetic Algorithm is given. Third, we constructed a two-level RFID Networks Planning model based on Distributive Decision-making method. According to the complexity of this two-level DDM model, a novel optimization algorithm based on System Intelligence is proposed to solve it.The simulation experiment indicates that the new SI method enhances the capacities including the adaptability, the robustness and the balance of the exploration and exploitation remarkably. Last, the multi-reader data fusion is studied in the system of RFID-based target locating and tracing system. We proposed a novel architecture for RFID reader data fusion based on fuzzy clustering method. By analyzing the basic biological foraging behaviors of honey bees, we proposed a novel Swarm Intelligence algorithm - Bee Swarm Optimization (BSO), which simulating the different individual behaviors and the interation pattern of the Bee colony. We then embed the BSO into the target tracing system as the basic algorithm for fuzzy clustering. The experiment results indicate that the proposed architecture and algorithms can filter the monitoring data error of RFID readers and enhances the precision of the target locating and tracing obviously.
Language中文
Contribution Rank1
Document Type学位论文
Identifierhttp://ir.sia.cn/handle/173321/408
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
陈瀚宁. RFID系统优化模型与智能算法研究[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(40369KB) 开放获取--Application Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[陈瀚宁]'s Articles
Baidu academic
Similar articles in Baidu academic
[陈瀚宁]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[陈瀚宁]'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.