RFID系统优化调度模型及算法研究
发布时间：2018-08-01 09:15
【摘要】：随着物联网技术的发展和泛在应用,作为感知层的RFID系统也发展为多个阅读器组成的近距离无线局域网络系统,实现广域范围内标签标识物品的互联、信息感知和共享。由于RFID系统本身的动态性和不确定性,系统的资源分配与优化调度问题一般是非线性、多目标、大规模的复杂系统优化问题。研究复杂环境下RFID系统的优化调度模型及其智能优化算法,寻求在资源受限约束条件下的系统运行优化调度方案,成为物联网技术领域研究的面临的重要课题。此类课题的研究,对提高现有RFID系统的运行效率、服务质量以及自动化、智能化、协同化水平,推动物联网领域技术的进步,具有重要的理论意义和应用价值。论文针对目前RFID系统运行调度中普遍存在的优化问题,在对国内外相关研究综述的基础上,分析了现有系统的组成及运行调度模式,推导了阅读器识别范围的计算模型,确定了系统优化运行的边界参数,建立了多阅读器优化调度模型及求解模型的B细胞优化算法,设计了识别大量被动标签的DFSA时隙段识别算法以及多阅读器协作的分组标签识别算法,最后设计了一个基于情景感知的RFID系统框架,在MATLAB环境下,对不同标准下的RFID系统场景进行了仿真实验,验证了所提模型及算法的有效性和实用性。论文研究内容融合物联网技术、信息科学、计算机科学、自动化科学、运筹学、概率统计学、进化计算以及管理科学等多个学科的交叉领域。论文的主要创新工作包括以下内容:(1)针对现有rfid系统内多阅读器环境(mre)和稠密阅读器环境(dre)中,阅读器之间存在的干扰和碰撞问题,推导了基于信号干扰噪声比(sinr)的阅读器识别范围计算模型,根据阅读器功率和频道资源确定了产生阅读器-标签碰撞(rtc)问题的临界距离。在此基础上,针对多阅读器共享时隙和频道产生的资源竞争问题,把多阅读器调度分解为对孤立阅读器节点、多阅读器节点和稠密阅读器节点的优化调度问题,提出了资源分配的约束条件以及保证调度公平性的约束条件,建立了多阅读器调度的动态规划模型。该模型符合rfid标准和协议,有效降低了阅读器-阅读器碰撞(rrc)问题,避免了rtc问题;(2)提出了基于几何概率分布函数求解模型的b细胞优化算法(gd-oa-bcell)以及考虑调度公平性的bgd-oa-bcell算法,将动态规划模型按照稠密阅读器节点和非稠密阅读器节点细分成底层目标函数和顶层目标函数,以全局识别范围最大和工作阅读器最多为目标并统一作为抗原,将每一套可行调度方案作为抗体,根据亲和度值划分为优秀个体和普通个体,基于几何概率分布函数分别对普通个体的稠密阅读器节点基因段提出了底层区域变异算子,对优秀个体的非稠密阅读器节点基因段提出了顶层多点变异算子,加快了算法的收敛速度,提高了抗体的多样性,避免了局部最优。gd-oa-bcell算法得到的优化调度方案,既有效避免了rrc和rtc问题,又保证了公平性,以调度多阅读器协同、并行工作。bgd-oa-bcell在此基础上,具有更好的公平性和实用性;(3)提出了基于dfsa框架的时隙采样极大似然函数标签数估计方法以及相应的标签识别算法(sia),针对sia算法帧调整不稳定的问题,进一步设计了时隙段采样帧调整算法(SS-ASF),以降低TTC问题。仿真实验表明,该标签数估计方法在标签数较大时依然能够保持理想的估计误差,SS-ASF在SIA基础上进一步提高时隙利用率和标签识别速度,同时显著减少帧调整次数,实现了阅读器在标签识别过程的高效性;(4)针对RFID系统中多阅读器识别大量被动标签的需求,提出了基于标签分组的标签识别算法(G-SIA),并进一步设计了分组多级随机数标签识别算法(GM-IP)以及多阅读器协同并行工作方式,采用三种负载均衡策略,解决算法中的负载不均衡问题。仿真实验表明,G-SIA和GM-IP能够有效处理数量规模较大的标签,显著提高时隙利用率,实现了多阅读器协同并行识别过程。在上述研究的基础上,设计了基于情景感知的RFID系统自优化调度子系统(AODS-Bcell)框架,并在MATLAB下开发了仿真程序,对符合EPCGlobal C1 G2标准和GB/T 29768-2013标准规范下的RFID系统场景进行了仿真实验,验证了所提模型和算法的有效性,以及与RFID系统标准的兼容性。AODS-Bcell以已有优化调度方案为基础,结合情景感知参数,可进一步实现自优化过程。研究成果对现有RFID系统优化调度及规划设计具有重要的借鉴作用和实际应用价值。
[Abstract]:With the development and widespread application of the Internet of things technology, the RFID system, as a perceptual layer, also develops into a wireless local network system composed of multiple readers, which realizes the interconnection of labels identifying items in the wide area, information perception and sharing. Because of the dynamic and uncertainty of the RFID system itself, the resource allocation and optimization of the system The degree problem is generally nonlinear, multi-objective and large-scale complex system optimization problem. The optimization scheduling model of RFID system and its intelligent optimization algorithm under the complex environment are studied, and the optimal scheduling scheme for system operation under the constraint of resource constraints is sought, and it becomes an important subject in the research of the technology domain of the Internet of things. In order to improve the operating efficiency of the existing RFID system, the quality of service, the automation, the intelligence, the level of cooperation and the advancement of the technology in the field of animal networking, it has important theoretical significance and application value. This paper, based on the general optimization problems of the current RFID system operation and scheduling, is based on the summary of related research at home and abroad. The composition and operation scheduling mode of the existing system are analyzed, the calculation model of the reader recognition range is derived, the boundary parameters of the optimized operation of the system are determined, the B cell optimization algorithm of the multi reader optimization scheduling model and the solution model is established, and the DFSA time slot recognition algorithm for identifying a large number of passive tags and the multi reader association are designed. In the end, a RFID system framework based on situational awareness is designed. Under the environment of MATLAB, the simulation experiments of the RFID system scenes under different standards are carried out to verify the validity and practicability of the proposed model and algorithm. The research content is integrated with the technology of Internet of things, information science, computer science, and automation. The main innovations of this paper are as follows: (1) the interference and collision between the readers in the existing RFID system and the dense reader environment (MRE) and the dense reader environment (DRE), and the derivation of the signal stem based on the signal stem are derived. According to the reader power and channel resources, the critical distance of the reader - label collision (RTC) problem is determined according to the reader power and channel resources. On this basis, the multi reader scheduling is decomposed to the isolated reader node for the sharing of time slot and channel resource competition in the multi reader. The optimal scheduling problem of reader node and dense reader node is proposed. Constraints of resource allocation and constraints of scheduling fairness are proposed. A dynamic programming model for multiple reader scheduling is established. The model conforms to RFID standards and protocols, effectively reduces the reader reader collision (RRC) problem and avoids the RTC problem. (2) (2) The B cell optimization algorithm (gd-oa-bcell) based on the geometric probability distribution function (geometric probability distribution function) and the bgd-oa-bcell algorithm considering scheduling fairness are presented. The dynamic programming model is subdivided into the underlying objective function and the top target function according to the dense reader node and the non dense reader node, and the global recognition range is the largest and the work is read. The device is the target and unified as an antigen. Each set of feasible scheduling schemes is used as an antibody. According to the affinity value, it is divided into excellent individuals and ordinary individuals. Based on the geometric probability distribution function, the underlying region mutation operator is proposed for the dense reader node gene segment of the ordinary individual, and the non dense reader node of the outstanding individual is used. The gene segment proposes the top-level multipoint mutation operator, which speeds up the convergence speed of the algorithm, improves the diversity of the antibody and avoids the optimal scheduling scheme obtained by the local optimal.Gd-oa-bcell algorithm. It not only effectively avoids the RRC and RTC problems, but also guarantees the fairness to dispatch the multi reader synergy and the parallel work.Bgd-oa-bcell. It has better fairness and practicability; (3) a method of estimation of the maximum likelihood function tag number of time slot sampling and the corresponding label recognition algorithm (SIA) based on the DFSA framework are proposed. In view of the problem of the instability of the frame adjustment in the SIA algorithm, a time slot sampling frame adjustment algorithm (SS-ASF) is designed to reduce the TTC problem. Simulation experiments show that the label is the standard. The number estimation method can still maintain the ideal estimation error when the number of tags is large. SS-ASF can further improve the time slot utilization and label recognition speed on the basis of SIA, and reduce the number of frame adjustment significantly, and realize the efficiency of the reader in the label recognition process. (4) a large number of passive tags are identified for the multi reader in the RFID system. The label recognition algorithm based on label packet (G-SIA) is proposed, and the packet multilevel random number label recognition algorithm (GM-IP) and multi reader cooperative parallel working mode are designed. Three load balancing strategies are used to solve the load imbalance in the algorithm. The simulation experiment shows that the number of G-SIA and GM-IP can effectively deal with the number. On the basis of the above research, the framework of RFID system self optimized scheduling subsystem (AODS-Bcell) is designed on the basis of the above research, and the simulation process is developed under MATLAB, which conforms to the EPCGlobal C1 G2 standard and the GB/T 29768-2013 standard. The simulation experiment of RFID system scene under the quasi standard has been carried out to verify the validity of the proposed model and algorithm, and the compatibility with the standard of the RFID system..AODS-Bcell, based on the existing optimal scheduling scheme, combined with the situational awareness parameters, can further realize the self optimization process. The research results are the optimal scheduling and planning of the existing RFID system. It has important reference function and practical application value.
【学位授予单位】：太原理工大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：TP391.44;TN929.5
本文编号：2157102
[Abstract]:With the development and widespread application of the Internet of things technology, the RFID system, as a perceptual layer, also develops into a wireless local network system composed of multiple readers, which realizes the interconnection of labels identifying items in the wide area, information perception and sharing. Because of the dynamic and uncertainty of the RFID system itself, the resource allocation and optimization of the system The degree problem is generally nonlinear, multi-objective and large-scale complex system optimization problem. The optimization scheduling model of RFID system and its intelligent optimization algorithm under the complex environment are studied, and the optimal scheduling scheme for system operation under the constraint of resource constraints is sought, and it becomes an important subject in the research of the technology domain of the Internet of things. In order to improve the operating efficiency of the existing RFID system, the quality of service, the automation, the intelligence, the level of cooperation and the advancement of the technology in the field of animal networking, it has important theoretical significance and application value. This paper, based on the general optimization problems of the current RFID system operation and scheduling, is based on the summary of related research at home and abroad. The composition and operation scheduling mode of the existing system are analyzed, the calculation model of the reader recognition range is derived, the boundary parameters of the optimized operation of the system are determined, the B cell optimization algorithm of the multi reader optimization scheduling model and the solution model is established, and the DFSA time slot recognition algorithm for identifying a large number of passive tags and the multi reader association are designed. In the end, a RFID system framework based on situational awareness is designed. Under the environment of MATLAB, the simulation experiments of the RFID system scenes under different standards are carried out to verify the validity and practicability of the proposed model and algorithm. The research content is integrated with the technology of Internet of things, information science, computer science, and automation. The main innovations of this paper are as follows: (1) the interference and collision between the readers in the existing RFID system and the dense reader environment (MRE) and the dense reader environment (DRE), and the derivation of the signal stem based on the signal stem are derived. According to the reader power and channel resources, the critical distance of the reader - label collision (RTC) problem is determined according to the reader power and channel resources. On this basis, the multi reader scheduling is decomposed to the isolated reader node for the sharing of time slot and channel resource competition in the multi reader. The optimal scheduling problem of reader node and dense reader node is proposed. Constraints of resource allocation and constraints of scheduling fairness are proposed. A dynamic programming model for multiple reader scheduling is established. The model conforms to RFID standards and protocols, effectively reduces the reader reader collision (RRC) problem and avoids the RTC problem. (2) (2) The B cell optimization algorithm (gd-oa-bcell) based on the geometric probability distribution function (geometric probability distribution function) and the bgd-oa-bcell algorithm considering scheduling fairness are presented. The dynamic programming model is subdivided into the underlying objective function and the top target function according to the dense reader node and the non dense reader node, and the global recognition range is the largest and the work is read. The device is the target and unified as an antigen. Each set of feasible scheduling schemes is used as an antibody. According to the affinity value, it is divided into excellent individuals and ordinary individuals. Based on the geometric probability distribution function, the underlying region mutation operator is proposed for the dense reader node gene segment of the ordinary individual, and the non dense reader node of the outstanding individual is used. The gene segment proposes the top-level multipoint mutation operator, which speeds up the convergence speed of the algorithm, improves the diversity of the antibody and avoids the optimal scheduling scheme obtained by the local optimal.Gd-oa-bcell algorithm. It not only effectively avoids the RRC and RTC problems, but also guarantees the fairness to dispatch the multi reader synergy and the parallel work.Bgd-oa-bcell. It has better fairness and practicability; (3) a method of estimation of the maximum likelihood function tag number of time slot sampling and the corresponding label recognition algorithm (SIA) based on the DFSA framework are proposed. In view of the problem of the instability of the frame adjustment in the SIA algorithm, a time slot sampling frame adjustment algorithm (SS-ASF) is designed to reduce the TTC problem. Simulation experiments show that the label is the standard. The number estimation method can still maintain the ideal estimation error when the number of tags is large. SS-ASF can further improve the time slot utilization and label recognition speed on the basis of SIA, and reduce the number of frame adjustment significantly, and realize the efficiency of the reader in the label recognition process. (4) a large number of passive tags are identified for the multi reader in the RFID system. The label recognition algorithm based on label packet (G-SIA) is proposed, and the packet multilevel random number label recognition algorithm (GM-IP) and multi reader cooperative parallel working mode are designed. Three load balancing strategies are used to solve the load imbalance in the algorithm. The simulation experiment shows that the number of G-SIA and GM-IP can effectively deal with the number. On the basis of the above research, the framework of RFID system self optimized scheduling subsystem (AODS-Bcell) is designed on the basis of the above research, and the simulation process is developed under MATLAB, which conforms to the EPCGlobal C1 G2 standard and the GB/T 29768-2013 standard. The simulation experiment of RFID system scene under the quasi standard has been carried out to verify the validity of the proposed model and algorithm, and the compatibility with the standard of the RFID system..AODS-Bcell, based on the existing optimal scheduling scheme, combined with the situational awareness parameters, can further realize the self optimization process. The research results are the optimal scheduling and planning of the existing RFID system. It has important reference function and practical application value.
【学位授予单位】：太原理工大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：TP391.44;TN929.5
【参考文献】
相关期刊论文 前10条
1 付钰;钱志鸿;孟婕;王雪;;基于连续时隙预测的帧时隙Aloha防碰撞算法[J];电子学报;2016年09期
2 张小红;胡应梦;;分组自适应分配时隙的RFID防碰撞算法研究[J];电子学报;2016年06期
3 蒋毅;张若南;程伟;李彬;;一种基于标签分布距离的RFID碰撞避免算法[J];西北工业大学学报;2016年02期
4 杜永兴;白文浩;李宝山;;RFID系统动态帧时隙ALOHA算法的改进[J];高技术通讯;2015年03期
5 刘晓慧;钱志鸿;赵雁航;郭雨齐;;一种自适应的无线射频识别系统中的防冲突算法(英文)[J];中国通信;2014年07期
6 史长琼;肖瑞强;吴丹;;一种改进的ALOHA防碰撞算法[J];计算机工程与应用;2016年04期
7 陈毅红;冯全源;;按需时隙分配RFID防碰撞协议研究[J];电子学报;2014年02期
8 庞宇;彭琦;林金朝;周前能;李国全;吴玮;;基于分组动态帧时隙的射频识别防碰撞算法[J];物理学报;2013年14期
9 张光山;张烁;张有光;;基于随机时隙的RFID读写器防冲突方法[J];北京航空航天大学学报;2013年06期
10 陈颖;张福洪;;RFID传感网络中多阅读器碰撞算法的研究[J];传感技术学报;2010年02期
，本文编号：2157102
本文链接：https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/2157102.html
