动态RFID系统中海量标签防碰撞技术研究

As one of the key technologies of internet of things (IoT) networks, radio frequency identification (RFID) has been widely applied in various areas, such as logistics, healthcare, manufacture, and security etc.. However, with the rapid expansion of the scale of the IoT networks, collisions caused by thousands of RFID tags’ simultaneous access increases sharply, resulting in a seriously decreased bandwidth utilization and a significantly increased identification delay. Moreover, in dynamic environments, it is also of great importance to effectively estimate and quickly identify the newly arrived and left tags. Focusing on the large-scale tag anti-collision technologies for dynamic RFID systems, this proposal consists of the following aspects: Firstly, we explore an effective strategy to detect the state of each slot, and propose a cross-layer based tree-splitting tag number estimate algorithm to improve the time efficiency and the estimation accuracy. Secondly, we design the idle slot-detecting and early-breaking strategies, and optimize the structures of the multi-branch and dynamic branch tree algorithms, which increases the time efficiency of the unknown tag identification process. Thirdly, we investigate the effective method of distinguishing the known and unknown tags, and propose an early-breaking and dual-confirmation algorithm to improve the reliability and time efficiency of the missing tag identification process. This proposal aims at providing a basic theory and method for the efficient tag identification in dynamic RFID systems and providing technical support and reservation to improve the intelligence of IoT networks.

射频识别（RFID）技术作为物联网领域的核心技术之一，广泛应用于物流、医疗、制造、安全等各个领域。然而，随着物联网规模的迅速扩大，海量RFID标签接入信号之间的碰撞急剧增多，导致网络信道带宽利用率严重下降、识别延迟显著增大。同时，在动态条件下，如何对新加入和被移出的海量标签进行可靠估计和快速识别，也是系统性能提升的关键。针对上述问题，本项目紧密围绕动态RFID系统中海量标签防碰撞技术展开，主要研究内容包括：①探索标签时隙选择信息的有效检测机制，提出跨层分支树标签数量估计方案，提升标签估计性能。②设计空闲时隙检测和帧中断分层策略，优化多分支树和动态分支树结构，提高未知标签识别效率。③研究已知和未知标签的有效区分方法，提出帧中断静默和双重确认识别方案，提高丢失标签识别的可靠性和时间效率。本项目旨在为动态RFID系统中的高效标签识别提供基础理论和方法，为物联网智慧化程度的提升提供技术支持和储备。

随着物联网的发展，RFID射频识别系统被广泛应用于物流、仓储、农业、医药以及智慧工厂等各个领域。在海量RFID系统中，标签接入信号碰撞严重，信道利用率低，识别延迟大。特别是在动态环境下，大量贴有标签的物品频繁加入和移出阅读器的读取范围，造成标签集合动态变化，增加了识别难度。本项目紧密围绕动态RFID系统中海量标签防碰撞技术展开研究，主要研究内容包括：①提出了基于比特检测跨层估计的多分支树海量标签防碰撞协议，结合信道分析设计基于比特检测的快速标签数量估计，多分支树标签识别以及哈希静默方法，提高标签数量估计准确度和标签识别可靠性，缩短识别延迟；②提出了基于随机采样查询估计的跨层分支树海量标签防碰撞协议，结合时帧结构与查询过程设计基于前缀匹配的随机查询标签数量估计方法，通过利用估计结果优化冲突树分组进程，提高识别效率；③提出了基于动态查询的多分支树快速移动标签高效防碰撞协议，利用标签流速统计结果设计多轮准入控制方法，优化动态标签识别机制，并通过动态查询多分支树方法减少碰撞时隙，减低标签漏读率；④提出了基于帧中断静默和双重确认的海量已知标签识别协议，通过帧中断点优化、多哈希映射以及标签紧凑响应等方法，提高未知标签数量估计准确度，优化时隙利用率，提高已知标签识别可靠性。本项目的研究可以有效提升动态RFID系统中的标签识别效率，为物联网的普及和应用提供理论基础和技术储备，在构建智慧城市、智能农场、智慧工厂、物品追踪溯源以及物资管理等方面具有重要的理论和实践指导意义。