物联网中一种基于树型结构的RFID防碰撞算法研究

发布时间：2018-01-12 07:32

本文关键词：物联网中一种基于树型结构的RFID防碰撞算法研究　出处：《吉林大学》2017年硕士论文　论文类型：学位论文

【摘要】：在信息科技高速发展的今天,物联网(Internet of Things,IOT)技术的重要性正越来越凸显出来。射频识别技术(Radio Frequency Identification,RFID)作为物联网感知层采集数据信息的关键技术,凭借快速读取、多目标识别、成本低、数据安全性高、非视距等优点,使其在未来应用会更加广泛。在未来物联网应用环境下,所需感知的底层数据必然是海量大规模的,而通信信道资源是有限的,当多个标签争用同一无线信道与阅读器通信时,不可避免地会产生数据碰撞问题,影响系统的识别速度与识别效率。标签数据碰撞问题的解决对RFID系统整体性能的提高至关重要,尤其是在未来物联网应用的大规模标签数据环境下,解决这一问题的重要性显得尤为突出。因此,需要建立有效且高效的防碰撞机制,用以解决多标签识别的碰撞问题,这也成为国内外学者的研究热点方向。RFID系统标签防碰撞算法主要有两大类,一类是基于ALOHA的不确定性防碰撞算法,一类是基于树型结构的确定性防碰撞算法。本文主要研究确定性的树型结构防碰撞算法,并针对现有的树型结构防碰撞算法在大规模标签识别时存在的搜索深度过深、碰撞时隙数与查询时隙数过多的问题,提出了一种新型有效算法---增强型锁位后退防碰撞算法(Enhanced algorithm based on BLBO,EBLBO),具体在以下三方面进行了改进:1.在锁定标签碰撞序列后,提出每三位碰撞位一识别的方式,即八叉搜索识别方式,来有效减少碰撞时隙数与标签识别延时,提高了系统吞吐率;2.提出一种碰撞前缀预测的方法来有效地消除由于采用每三位碰撞位一识别所带来的空闲时隙增加的问题,避免空闲时隙所带来的系统标签识别性能下降;3.在搜索树算法中引入堆栈思想,即在阅读器中维护一个碰撞堆栈,将预测出来的确定存在的碰撞前缀存储于碰撞堆栈中,查询时依次弹出栈首碰撞前缀进行查询,避免回到根节点处查询所带来的冗余。然后对所提新算法EBLBO在阅读器的查询次数、传输数据量、标签平均识别延时及系统吞吐率等性能指标上,进行了理论推导及仿真实验。实验结果表明,在大规模标签识别环境下,新算法通过这三点改进,较原锁位后退防碰撞算法BLBO,阅读器查询次数减少近16.6%,系统吞吐率提高近10%,传输数据量减少近48%,标签平均识别延时减少近1.36ms;有效降低了碰撞时隙数与查询时隙数,进一步减少了传输数据量及标签识别延时,系统吞吐率及系统整体识读性能得到有效提高,证明了所提新算法的有效性与优越性。最后分析与讨论了碰撞前缀预测位数对系统标签识别性能的影响,并给出碰撞前缀预测位数与阅读器查询次数、传输数据量、标签平均识别延时及系统吞吐率的关系曲线,论证出使系统标签识别性能达到最优时的碰撞前缀预测位数。
[Abstract]:With the rapid development of information technology, Internet of things (Internet of things). The importance of IOTs is becoming more and more important. Radio Frequency Identification is a radio frequency identification technology. RFIDs as the key technology of collecting data in the perceptual layer of the Internet of things, with the advantages of fast reading, multi-target recognition, low cost, high data security, non-line-of-sight and so on. In the future Internet of things application environment, the need to perceive the underlying data must be massive and large-scale, and the communication channel resources are limited. When multiple tags compete for the same wireless channel to communicate with the reader, the problem of data collision will inevitably occur. The solution of tag data collision problem is very important to improve the overall performance of RFID system, especially in the large scale label data environment of Internet of things applications in the future. The importance of solving this problem is particularly prominent. Therefore, it is necessary to establish an effective and efficient anti-collision mechanism to solve the collision problem of multi-label recognition. This has also become a hot research direction of domestic and foreign scholars. There are two main categories of anti-collision algorithms for RFID system tags. One is the anti-collision algorithm based on ALOHA uncertainty. One is the deterministic anti-collision algorithm based on tree structure. And the existing tree structure anti-collision algorithm in large-scale label recognition in the search depth is too deep, collision time slot and query slot number too many problems. An enhanced algorithm based on BLBOLBON (enhanced algorithm based on BLBOLBON) is proposed. After locking the tag collision sequence, an octal search method is proposed to effectively reduce the number of collision slots and the tag recognition delay. The system throughput is improved; 2. A collision prefix prediction method is proposed to effectively eliminate the problem of increasing the free time slot caused by every three bit collision bit recognition, and to avoid the degradation of the system label recognition performance caused by the free time slot. 3. The stack idea is introduced into the search tree algorithm, that is, to maintain a collision stack in the reader, and store the predicted collision prefix in the collision stack. In order to avoid the redundancy caused by the query at the root node, the first collision prefix of the stack is popped up when querying, and then the number of times of query in the reader of the proposed new algorithm EBLBO, the amount of data is transmitted. The theoretical derivation and simulation results show that the new algorithm is improved by these three methods in the large-scale label recognition environment. Compared with the original anti-collision algorithm BLBO, the reader query times are reduced by 16.6, the throughput of the system is increased by nearly 10 percent, and the amount of data transmitted is reduced by nearly 48%. The average recognition delay was reduced by 1.36 Ms. It can effectively reduce the number of collision slots and the number of query slots, further reduce the amount of data transferred and the tag recognition delay, and improve the system throughput and the overall performance of the system effectively. The effectiveness and superiority of the proposed algorithm are proved. Finally, the influence of the predicted bit number of collision prefix on the recognition performance of the system label is analyzed and discussed, and the prediction bits of collision prefix and the number of reader queries are given. The relation curves of transmission data, tag average recognition delay and system throughput are discussed, and the predicted bits of collision prefix when the system label recognition performance reaches the optimal level is demonstrated.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.44;TN929.5

【相似文献】