云南地区RFID物流供应链的仿生定位感知MAC协议研究

Supply chain (SC) based on passive radio frequency identification (RFID) is an improtant application field in Internet of Things, and medium access control (MAC) protocol for passive RFID tag is a key technology in SC based on passive RFID. Logistics in Yunnan region has the following features: too many identify items, complex physical property of identify items and simple identifycation systems. Since conventional MAC protocols consider passive RFID as a random access system, require blind tag number estimation and hardly consider capture effec, the convential protocols applied to the region have the disadvantages of longer identification delay, higher computational complexity and more identification ommision. For the disadvantages, this proposal will paropse a MAC protocol, and investigate a tag information transmission mode and a location perception mode based on bionic echolocation effect. In this proposal, we consider an RFID system as a un-random system by dividing tags in SC into several moved tag sets, and identify the tag by the echolocation information. First, the proposed MAC protocol will adopt simulated-ultrasonic pilots to collect a vector of idle, successful and collision information; second, the protocol will unblindly process the vector, and use an adaptive algorithm and a nested binary tree algorithm to percept the tag echolocation information; finaly, a reader can identify tags by the percepted information. From the methods above, this proposed protocol may reduce idientification delay, lower computational complexity and avoid identification ommision.

无源射频识别(RFID)的物流供应链是物联网的重要应用领域之一，而其标签的MAC协议又是无源RFID物流供应链的支撑技术之一。由于云南地区物流具有识别数量多、识别对象物理性能复杂和识别系统应简单的特点，而传统的MAC协议把无源RFID作为随机多址系统、需对标签数做盲估计且较少考虑捕获效应，因此用于该地区存在识别时间较长、计算复杂度较高和标签易漏读等问题。针对传统MAC协议的这些问题，本项目拟构建一种仿超声波回声定位效应的标签信息传输模式及定位信息感知模式。首先把供应链中运动标签集合作为基本识别单位，实现非随机多址系统，然后借鉴动物对定位信息的获取与处理的反馈机理，采用仿超声导频序列获取空闲、成功和冲突的多维矢量信息，再以非盲的原则对该矢量信息进行处理，利用自适应学习和嵌入式二进制树算法来感知定位信息以实现对目标标签的识别，从而达到减少识别时间、降低复杂度和防止标签漏读的目的。

物流供应链管理是无源射频识别(RFID)系统的重要应用领域之一，而研究RFID标签的MAC协议对如何在物流供应链中更好地完成标签识别具有重要意义。在物流供应链中，标签识别环境较为复杂，捕获效应的发生是一普遍现象，而且标签识别数量也会增加。传统的MAC协议应用在物流供应链环境中，存在识别时间较长、计算复杂度较高和标签易漏读等问题。针对这些问题，本项目研究了以下内容：（1）面向无源RFID运动标签集合的识别研究，（2）仿超声回声定位效应的多标签识别研究，（3）捕获效应中嵌入式二进制树标签分配策略；得到了以下的研究成果：（1）标签集合与回声定位的自适应二进制树标签识别方法与理论，（2）捕获效应下的最优标签识别效率，（3）稠密标签环境下的RFID标签识别方法与理论。从实验结果看，与传统方法相比，本成果提出的理论和方法避免了标签漏读，减少了识别时间，并具有较低的计算复杂度。本研究成果的价值在于，将其提出的理论和方法应用到捕获效应的环境下，不会产生标签漏读，还可提高标签的识别效率，同时，当应用在大规模标签场合，可加快标签识别过程，而且本成果所得到的算法具有较低的计算复杂度，能应用到简易且高效的RFID识别系统中。