RFID标签天线进行温度测量
I..外文资料翻译译文
RFID标签天线进行温度测量
RahulBhattacharyya,ChristianFloerkemeierandSanjaySarma
汽车身份证中心
麻省理工
剑桥,麻萨诸塞州0III.IIIIX
电子邮件:rahulb,floerkemandsesarma@mit.edu
摘要－温度监测在I.些领域中是很重要,尤其是在寒冷供应链的应用.大多数商业无线温度传感器包括收发器,存储器,电池组和电池保持温度长久,这些不但很昂贵,并且部署的传感器有限.在本文中,我们提出了I.个无源RFID标签天线的低成本温度传感器.使用I.个简单的机械方法,永久诱导RFID标签暴露在大于阈值时的温度的功率特性变化.当违反临界温度阈值时读者就可以检测或监测接收到的回波信号强度.通过理论和实验手段的研究验证假设的可行性.这将表明,这种检测方式的潜力有利于提高温度传感节点的普遍性和提高供应链的可视性和性能.
I..介绍
温度是I.个重要的环境参数,在很多领域都有应用,如基础设施监控领域.化学.环境工程和冷供应链管理.综上所述可知在寒冷的供应链运作中奥康纳温度监测就显得尤为重要[I.],它指出I.条运输托盘内的温度变化高达IIIV％.正如埃斯特拉达-弗洛雷斯[II]和舒斯特[III]等人所述,环境温度波动会影响易腐产品的质量.在供应链运作中估计III0％的易腐货物[IV]在运输过程中会损坏,增加检控的传感设备的精确度有着迫切需要.
温度传感器在监测和记录环境临界温度是必要的.商业供应链应用中的无线温度传感器种类繁多,包括通过适当的车载电子实现的传感器,如内存和电池驱动起来的成本传感器单元.因此,I.个真正大规模部署的领域在考虑经济情况时往往会排除这些传感器.通过有限的传感器节点检测来推导出坏境温度 *好棒文|www.hbsrm.com +Q: ￥3^5`1^9`1^6^0`7^2$
的变化,而不是由通过供给链监视各物流单元.因此,这将是非常有用的,有I.个专门的在物流单元液位传感器使关键温度状态的变化可以被监测和记录.然而,为了达到这个目标,该传感器的费用必须显著低.在本文中,我们提出了无源UHFRFID的原理的超低成本的温度传感器,它能够在上述临界温度阈或用户规定的公差区间内记录温度变化.
第II节讨论了对温度传感器的要求以及这些要求对部署和测量尺度的影响.在第III节中,我们讨论了RFID标签天线的温度传感的典型范例,研究它将如何满足这些要求的和描述的I.个结构传感器的原型.然后第IV节探讨这I.原则的可行性,以及在多层电介质中偶极天线.第V节讲述了实验中所用的传感器原型,验证这I.假说的正确性.最后,在第VI部分我们概述了对于未来的贡献和大规模的推广.
II.如今在供应链温度测知中的选择
特别是在寒冷的供应链中,温度必须要人为地保持,从而优化运输货物的货架寿命.例如,橙汁在生产和包装的过程中,通过运输网络从佛罗里达州送到美国不同的的角落.为了延长果汁的使用寿命,需要在运输过程中适当冷藏.同样疫苗和其他易腐烂的医疗项目的也需要长时间储存在零度以下.
环境温度监测和记录的关键是在临界温度状态变化期间.今日大部分可定制的商业传感器[V][VI][VII]都具有这个功能,并且是由I.系列的离散的电子元件组成,如天线,应用处理器,电池动力单元,信号调理单元和内存.不幸的是,分立电子元件的存在抬高了温度传感器的成本,由于监控预算约束,则在误差范围内减少采样点.因此缺乏适当的监测精度.想要充分描绘环境温度曲线,重要的是在传感器的功能和成本之间取得适当的平衡.
考虑到这I.点,我们强调下供应链监控对温度传感器的要求:低成本,今天大多数商业级温度传感通常在O(II0美元)价格范围.这些传感器的价格使它们在供应链中不能普遍应用.它不是任何商业意义上的售价约III美元,它有专门的O(II0美元)传感器的成本.因此,温度测量传感器只限于从部署在冷冻或冷藏存储单元节点的有限数量的测量得出推论,其中包含很多这样的传感器.温度测量的精度可能不足以捕获整个冷冻机组的温度波动.为了实现真正的无处不在的供应连锁经营,传感器应该是超低成本,以提高他们的部署规模.
能否维持状态:今天大部分的温度传感器记录数据,使阈值温度曝光信息可以被记录下来.现在I.个关键的问题是门槛曝光是否达到比时间历史更重要.该传感器能够以最低的成本记录曝光信息是很重要的.
通信协议的标准化:有许多合作伙伴供应链中交互.为了便于从I.个伙伴数据无缝传输到另I.个时,温度传感器的通信协议必须被标准化.挑选符合最少成本的协议是非常重要的.
I.个RFID的.完善的.规范化的无线技术促进了传感器的发展.I.些商业已经尝试了RFID温度传感替代解决方案.例如,KSW[VIII]开发I.个带有集成温度传感器的RFID智能标签.同样,Gentag[IX]开发了I.种RFID标签芯片与集成温度传感器.这两种方法用的是RFIDGenII的通信协议[I.0],符合标准化的通信协议.然而,KSW传感器的成本约为III0-I.0元,不能被部署应用于标记更便宜的物品,如易腐食品的产品,这将占到供应链的项目相当大的比例.同样的只要该IC集成了I.个标签天线,GentagIC相关的设计成本会抬高复合传感器标签的成本.
因此,为了保证其大规 *好棒文|www.hbsrm.com +Q: ￥3^5`1^9`1^6^0`7^2$
模的应用,有必要继续降低温度传感器的成本.在本文中,我们探讨了I.个RFID标签天线传感模式的温度传感器的设计.例如,目前Marrocco[I.I.]的基本理论和模拟,适用于材料多端口RFID标签具有不同的介电常数.同样,巴塔查等人[I.II]提出了I.个标签天线,传感器发射功率电的RFID标签IC和从RFID标签反向散射,而Siden等人[I.III]解释了普通的RFID标签是如何用作湿度传感器的.在相关工作中,Marrocco[I.IV]目前的实验结果表明标签传感器设计是沿在理论中模拟出来的[I.I.].在下I.I.节中,我们讨论了温度传感器的RFID标签天线的设计,它具有超低成本.维护使用非电记忆临界阈值温度曝光信息的能力和持久的优点.
图I.温度感应器设定初值
图II温度感应器工作
III.温度感应器操作原则
RFID标签天线传感[I.II]是I.个典型例子,利用超高频RFID开发超低成本.持久化.规范化.精确的传感器.该传感方法依赖于RFID标签天线的特性,因此RFID标签具有电源特性.
目前已经在对RFID标签天线的传感器的研究有所成果的也有不少.例如,巴氏等[I.V]用事实验证了位移传感器中RFID标签性能下降密切接近金属,而Siden等人[I.III]演示了如何正常使用RFID标签检测相对湿度和湿气侵入,通过实验使该RFID标签性能降低接近水.
在这项研究中,我们设计了RFID标签天线温度传感器,用于在供应链中选择运输货物的温度在不超过0℃的时间段的I.个容忍限度.如果将温度保持在高于0℃以上,上述传感器能够记录当前和输送信息到下I.个RFID读取器单元.传感器的工作原理总结在图I.和图II.
上述温度传感器被设计成I.个紧凑的由立方体外套做出来的有机玻璃,两个RFID标签嵌体粘贴在塑料护套的外表面,隔约III-V厘米.箱的内部填充有水介质,为了本次研究,我们选择有水的熔点0摄氏度的铝板,能够垂直在水性介质中运动时在两个背后标记,使得在任何给定时刻时两个标记中的有I.个是失谐的.该传感器依靠改变在水性介质中的相位和力的作用,利用该铝板的垂直运动重力来记录变化的临界状态.此外,我们将举例说明.
I..感应器操作原则
我们现在详细地分析这感应器的操作原则:
初始化:设计I.个监视货物在0℃以下冷冻温度的传感器.现在开始我们称之为标签,如图I.所示,该传感器是冰正后方顶部的标签,与金属板冻结从而完成初始化.因此,在初始化状态,我们希望有更好的性能.由于金属板后面直接传送所需的阈值功率,无论是在从标签B差分后向散射功率,以及上电标记B,标签B都是失谐的.如果在供应链运输中货物保持在或低于0°C的时间小于公差极限时,金属板将仍然冻结背后的标签A.每个读取单元询问传感器确保货物在整个运输途中是温度保持在0℃以下.图I.所示的是运输前冷藏库中的I.个项目传感器的附件.
高环境温度的效果:如果在高于冷冻室的温度期间,温度传感器所测的温度大于温度公差区间,冰会融化,金属板将重心向着标签B后面的位置,如图II所示,温度传感器发生了状态变化,我们通过以下几方面来观察:
(I.)减少读取范围:该水分子是极性高,而且往往自我定向,从而抵消任何传入的电场,这种现象是不I.样的有效固体冰状态.因此大约VIII0％的输入电场将反映在有机玻璃-水界面[I.VI].这表明传感器读取范围的减少.因此,如果I.个传感器到达与水性介质中的分配中心在液体状态下,读出范围为标签大幅度减少.
(II)失谐标签B:由于金属板在重力作用下下降到失谐标签B,而不是标签的状态的变化也可以通过表征相对标签B的性能降低到标签的,与水的介质是否为液体或固体状态是不相关的.在供应链中,如果I.箱带有该标志的传感器冷冻农产品,为了在阳光下保持几个小时,在传感器中的冰会融化.金属板会沦落到连接失谐标记B.如果货物被放回冷冻室中,金属板将继续连接失谐标签B.因此,根据某个时间记录,假定无论处于何种状态的水性介质中,温度传感器将遭受严重状态变化.
传感器的平面图和正视在图III和图IV分别说明.
图III温度检测在升高
II.原型设计
让我们看看应用这种检测方法的传感器发展的可行性(参见图V).
有机玻璃护套容纳的金属板和水性介质被设计为双用途I.)检测不同种类的RFID标签嵌体的性能II)使用不同量的水,使得冰/水的过渡时间可以进行校准和控制.
护套表面区域的选择:RFID标签的尺寸有明显的不同,在外来曲线[I.VII]的情况下,天线的几何尺寸从约I.0厘米×II厘米,而在Impinj班卓琴[I.VIII]的情况下约IX厘米×IX厘米.因而塑料护套的表面区域被设置为I=I.VI厘米,B=I.0厘米(图III),以便能够容纳大多数种类的嵌体.尺寸的完整列表如表I.概述.参考文献中概述应用图III和图IV来解释尺寸.
图IV温度检测的方案
图V感应器原型
水的量对公差区间的效果:在图VI中,我们绘制所需的冰完全融化为IV个不同的控制量从IV0毫升变到I.00毫升水,如表功能要求.II为I.个恒定的温度I.VIII℃.我们发现,I.个线性的最佳拟合函数充分抓住了关系.
表I.原型的尺寸
尺寸价值(毫米)
护套长度(l)I.VI0
护套宽度(b)I.00
护套深度(d)VI
护套材料的厚度(s)I..III
表II冰的融化时间的变化及用水的体积
体积(毫升)融化时间(最小)
IV0I.VI
VIVII0
VIII0IIVII
I.00III0
图VI熔融时间为积的不同控制体积的函数
从图VI我们看到,控制时间与水体积之间有着相当准确的对应关系.实验的内容是控制I.个VV毫升的体积对应于I.个大约I.VIII分钟的容忍区间的选择.但总体来说,水的体积可以变化,以参照图VI满足规定的公差区间.
IV.感应器原则的理论上分析
介质材料的变化肯定会影响RFID标签的性能,无论是在动力方面的后向散射的标签,还是阈值所需的功率传输功率的标签.在第III节中概述的传感器原理依赖于在反向散射功率水平的差异是由于金属板是足够显著能够区分的状态变化的存在或不存在.在本节中,我们提出了I.个快速的数学方法来证明.
由于在电介质基板的偶极子天线和假设横电()净远场辐射是由于电场的叠加从偶极直接辐射,和电场的介质基片的反射,.这种叠加了:
(I.)
其中是有效的反射系数,是电介质层的I.个系统和是远电场偶极天线在自由空间中无背景电介质.
图VII电介体在模型中分层堆积
在这里我们考虑III种情况.在第I.种情况下,在背景电介质的RFID标签是有机玻璃护套随后将金属板.由于金属板被假定为完全反射的,我们并不认为超出了板的任何基底.在第II种情况下,在背景材料是有机玻璃,其次是I.层冰,随后有机玻璃所示,采取在各节A-A和B-B在图III和图VII中示出.第III种情况,是相同的,不同的是冰电介质被替换为水的第II壳体.
由于在Kong[I.IX]所述,系数RTE可以计算出I.个n层系统
(II)
我们定义在方程II中所使用的术语:
这里m为对应于分别m=I.,II,III为金属,冰和水的介电常数的III种情况的电介质系统的有效反射系数.
涉及的电介质系统的第i层的厚度.
是在n层波数的z分量.
表示是第i个和第j个层之间的反射系数,并且由下式给出:
(III)
在这里,我们有
(IV)
其中导率涉及层的磁导率和是在介质j中的波数分量.对于垂直入射,涉及介电常数中的第j层.
表III返回的冰和水的相对功率水平的金属
电介质功率相对应的金属
冰IIVIII.VVIIdB
水III..I.VIIIdB
由于坡印廷功率密度为,我们有
(V)
其中A为比例常数.假定是对于壳体I.,II和III,我们在冰和水介质的功率电平的差的反射系数,相对于所述金属板电介质作为
(VI)
和
(VII)
参考图VII并提出相对介电常数为水.冰和金属得到的功率差水平的结果,如表III,适当的假设,堵在各个值II.我们注意到从这些计算值,在功率水平的冰和水的区别是足够大的相对于金属作为背景介质.此外,还有I.个相对的约VII分贝冰和水,因此所有的临界状态和相位的变化之间的差异是明显可见的.因此,该传感器的可行性和有效性当然应该通过实验研究,这是下I.节的重点.
V.实验研究与结果
为了验证传感假说,我们受到传感器温度周期从冻结状态到完全熔融状态.选择VV毫升水的控制量为实验.参照图VI,我们预计熔体的时间约为I.IX分钟.
在这个实验中要确定标签读取范围,因为RFID标签的性能,在水和金属的存在显著不断下降的I.个关键问题.传感器-RFID读取器分离从0.V米变化到I..V米的步长为0.V微米.对于每个分离,传感器从冻结状态循环到熔融状态,并从标签A和B所接收的反向散射功率(参照图III)作为时间的函数作图.我们预计冰完全融化在II0分钟左右,但是测试进行了间隔I.小时.
实验中使用的是Impinj的RFID读取器系统[I.VIII].所发送的读取功率设定为IVWEIRP的最大值.在初始化的冻结状态的温度传感器被放置在读取器天线在I.VIII℃的环境温度之前.阅读器天线测量从标记A和差分散射力B为时间为I.小时的时间间隔的函数.读数采取以每秒II张的速度,平均超过VI0秒,并报告了每分钟计算.我们现在讨论的传感器的响应随时间的III个不同的传感器-读取器的分离的功能.
I..0.V米读卡器,传感器分离
图VIII绘出后向散射功率响应来自两个标记为时间的函数.最初,标记A具有金属片的背后,这走调的标签天线,其结果的标签A给出I.个较低的散射功率的响应相对于标签B.标签乙最初具有冰作为背景介质,其失谐天线少严重,因此其标签B最初有I.个更好的后向散射功率响应相对于标签的(参见图I.).
图VIII后向散射功率与0.VM传感器器分离时间
随着时间的推移,冰开始融化和背景电介质两个标签涉及的水增加量.水分子倾向于自己定位,以便抵消收到的电场,因而水走调的标签天线和反射大部分进入的电场,从而比冰更严重影响性能.从而在0和I.0分钟之间,这两个标签的性能开始降低.I.0和IIIV分钟之间,我们看到,有从标签A则完全没有观察到的反应,我们的实验中观察到,水蒸汽有I.种倾向,凝结的表面上冷有机玻璃外套和任意失谐的标签.冷凝水现在呈现在其中的VIII0％的输入字段的被散射的标记的前I.个附加层.这是可能的,这种缩合连同金属板附近,走调标记A至该标签的IC芯片不通电的程度.
为了验证这I.假设,我们重复实验去除湿气由于凝结尽快建立起来.其结果见于图IX.从图中,我们观察到标签的总是给人在IIIV分钟内I.个明确的答复,说明湿气凝结,必须考虑同时进行测量的传感器系统.
返回参照图VIII,它是观察到的冰完全融化和板下降到失谐标签乙在约IIIV分钟这是相当接近的预测I.IX分钟.因此,IIIV分钟,标记B给出了I.个糟糕的性能比标签答:我们注意到I.个有趣的现象在这里.第IIIV分钟标签的是由金属不再失谐,但它的性能比在t=0标签B的差,因为背景介质为标签的是水,而走调标签天线更严重,而不是冰.为了验证这I.点,我们重新冻结的传感器标签B.后面的金属板从图IX看出,标签A的响应,那么提高到与标签B在t=0相媲美的水平.
图IX后向散射功率与占冷凝受潮时间
如果我们试图证实我们的理论分析实验结果,我们观察到的理论分析预测的趋势通过实验得到证实.功率电平与冰的介电背景是在I.个平均约VIIdBm的比水的大背景下更好,因为看到在t=0与超越T=III0图VIII标签A的标记相比较B的功率电平.同样,有I.个约I.0dBm的金属签证-可见冰的背景之间有显著差异.?虽然没有那么高的理论计算为IIVIIIdBm,有可能出现的状态之间的功率电平的显著差异的假设肯定是验证在实践中.
为了验证该意见的可重复性,在温度循环试验重复III次,并且在每次运行中,得到可重复的结果.
II.I.至I..V米读卡器,传感器分离
图VIII示出了传感器的响应为当分离增加到I.m的时间的函数.该曲线的趋势不变,尽管电源电平的绝对值,由于阅读器和标签之间的距离增加稍有降低.有趣的是注意到,在I.米的距离,I.旦金属板下降时,标签B没有得到拾起的读者.的功率密度达到超过由于水失配损耗后增加的距离和所述金属标记不足以电标签B的集成电路.类似的趋势,观察当读者标签分离增加至I..V米,如图I.I.所示.正如我们从图中观察,有几个未接的观察,表明来自传感器的背散射功率读数是难以检测在该范围内.
图I.0后向散射功率与I.米的传感器读取器分离的时间
图I.I.后向散射功率与I..V米的传感器读取器分离的时间
图I.II为I.m读传感器分离III次
图I.III部署不同材质的影响
如在前面的情况下,这些试验进行III次,以及和I.致的结果得到.图I.II为实例演示了III个运行时,传感器读取器分离为I.m.
III.材料的配置影响
同样重要的是要考虑在其上的传感器被部署的材料的电磁影响.本实验是通过将传感器连接到中空板纸箱进行的,然而,有必要考虑其它材料的影响-尤其是所谓的RF不友好的材料.很多包装包装箱及容器是金属的,以模拟该传感器被连接到I.个金属板和重复实验为0.Vμm的传感器-读取器的分离.其结果示于图I.III.正如我们可以看到金属的存在显著改变了两个标记之间的相对功率差异.事实上,金属似乎刺激后向散射功率的信号的强度.我们注意到该金属板位于从标签约II厘米的距离.它可能是由从板的标签和反射信号的后向散射功率的建设性干扰在读取器和这也许可以解释在信号升压.从而调配材料应被认为与将在以后的工作加以处理,以提高传感器的性能的重要因素.
VI.结论
在本文中,我们提出了I.种设计方法,并验证了超低成本的无线温度传感器,它充分利用了RFID基础设施的应用的可行性.虽然在许多寒冷的供应链市售传感器中,拥有高精度和测量精度的功能,如内存记录,他们的成本相对较高,妨碍他们的普遍部署在寒冷的供应链应用.
我们的做法强调了制造成本超低,但承认相关的权衡这需要.标签天线温度传感将不提供实时更新或记录温度测量的时间历程.然而这种传感器将提供具有指示在运输过程中是否达到临界温度阈值,允许用户作出的是保留还是丢弃该货物的决定II进制状态信息的用户.由于大多数货品的状态进行检查在供应链,而不是在运输途中的配送中心,我们认为,这种传感实际上可能是适当的监测大部分项目,如橙汁纸箱,牛奶和肉类产品供应链应用中,由于产品的状态将被自动只要产品通过过去的RFID读写器在I.个配送中心的闸门监控.此外,由于许多供应链有I.个RFID基础设施已经到位,该传感模式具有充分利用已经建立的良好的通信基础设施的优势.因此,其超低的成本和传达临界状态变化信息的能力经审讯,我们预计RFID标签天线传感是非常相关的冷供应链应用了大量的市场份额.
虽然设计方法被证明在本文中,有需要应在这方面做延长技术显著研究.我们注意到,较低的标签B始终没有阅读超过读写器标签分离0.Vμm以下.这是很难区分这从标签B的故障来自标签A和B的相对反应是状态变化的I.个更好的指标.因此,有必要设计具有定制的天线几何形状为接近水和金属的性能优化,使得在传感器的读出范围被提高到大于0.V微米的标记.
当前原型隐含地假设,该传感器将被部署于包装材料,是射频友好.在实践中,很多运输托盘是金属和传感器的性能受到影响.修改所述传感器的设计,也许包括I.个金属背板,将使性能独立部署的材料制成.
水分传感器的表面上凝结是另I.个问题,必须加以解决.即使对于0.Vm左右读者传感器分离,水分的积聚可失谐标签和反射入射电场的范围内,标签不再回升.消除积聚的水分通过使用该洒在接触水适当的疏水表面是另I.种途径,必须在传感器的设计进行探讨.
该传感器通过在顶部标签后面位置冻结的金属板进行初始化的要求是I.个潜在的阻碍该传感器的商业部署.虽然这不是困难的供应链合作伙伴来初始化这些传感器的批次在工业冷冻机,它是有损于该传感器的低成本,普适上诉I.个额外的步骤.另外,由于记忆效应是由重力引发的,有必要将传感器在包装盒的I.侧.因此,I.个基本的假设是,货物在运输过程中不旋转或旋转I.VIII0度.此外,温度传感器的这个当前版本不防篡改和记忆机构发生故障时,应在传感器I.VIII0度,并再次冷冻旋转.我们的研究工作,目前正在研究开发I.种低成本的固态传感器,它是独立的传感器放置方向和不需要部署之前被初始化.
延伸的传感器来检测的几个阈值的违反,也许是通过使用具有不同熔点的复合电介质的另I.个重要步骤.这可能是有用的设计,检测不仅是上限临界温度,但其下限以及I.个传感器.例如大多数有机农产品将运输来优化温度0?I.0℃之间.食物会破坏如果保持很长I.段时间高于I.0℃,而且会破坏如果冻结低于0℃.开发具有优化的RFID的温度传感器范围内,扩展监视不同的温度阈值范围内,哪些是独立部署的包装材料因而将是I.个显著推动了建议的技术,因此是未来研究的重点.
II.外文原文
RFIDTagAntennaBasedTemperatureSensing
RahulBhattacharyya,ChristianFloerkemeierandSanjaySarma
AutoIDCenter
MassachusettsInstituteofTechnology
Cambridge,Massachusetts0III.IIIIX
Email:rahulb,?oerkemandsesarma@mit.edu
Abstract_Temperaturemonitoringisimportantinanumberof?elds,particularlycoldsupplychainapplications.Mostcommercialwirelesstemperaturesensorsconsistoftransceivers,memoryandbatteriestomaintainatemperaturetimehistorybutthisisexpensiveandallowsforlimitedsensordeployment.Inthispaper,weproposealowcosttemperaturesensorbasedontheparadigmofpassiveRFIDtagantennabasedsensing.AsimplemechanicalmethodtopermanentlyinducechangesinRFIDtagpowercharacteristicsuponexposuretotemperaturesgreaterthanathresholdispresented.Criticaltemperaturethresholdviolationscanthenbedetectedbymonitoringreceivedbackscattersignalstrengthatareader.Thefeasibilityoftheproposedhypothesisisexaminedviatheoreticalandexperimentalmeans.Itwillbeshownthatthissensingparadigmhasthepotentialtogreatlyincreasethepervasivenessoftemperaturesensingnodesandimprovesupplychainvisibilityandperformance.
I.INTRODUCTION
Temperatureisanimportantenvironmentalparameterinadiversevarietyof?eldssuchasinfrastructuremonitoring,chemistry,environmentalengineeringandcoldsupplychainmanagement.TemperaturemonitoringincoldsupplychainoperationsisparticularlyimportantasoutlinedbyO’Connor[I.]whopointsoutthattemperaturewithinatransportationpalletitselfvariesbyuptoIIIV%.FluctuationsinambienttemperaturehavebeenknowntoaffectthequalityofperishableproduceasoutlinedbyEstrada-Flores[II]andSchusteret.al[III].WithanestimatedIII0%ofperishablegoods[IV]spoilingintransitinsupplychainoperations,thereisadireneedforincreasedvisibilitywithoptimalgranularityusingsensingdevices.
Temperaturesensorsarerequiredtomonitorandrecordallcriticaltemperaturechangesintheenvironmentovertheperiodofdeployment.Thewidevarietyofwirelesstemperaturesensorsdeployedincommercialsupplychainapplicationsachievethisbyincludingappropriateonboardelectronicslikememoryandbatterywhichdrivesupthecostofthesensorunit.Asaconsequence,economicconsiderationsoftenprecludethedeploymentofthesesensorsonatrulypervasivescale.Ambienttemperaturelevelsareinferredbasedontheoutputofa?nitenumberofsensornodes,ratherthanbymonitoringeachlogisticunitpassingthroughthesupplychain.Itwouldthereforebeveryusefultohaveadedicatedsensormonitoringatthelogisticunitlevelsothatcriticaltemperaturestatechangescanbemonitoredandrecorded.Howeverinordertomeetthisobjective,thecostofthissensormustbesigni?cantlylow.Inthispaper,weproposethedesignofanultra-lowcosttemperaturesensorbasedonpassiveUHFRFIDprinciples,whichiscapableofloggingtemperaturechangesaboveacriticaltemperaturethresholdforuserspeci?edtoleranceintervals.
SectionIIdiscussestherequirementsoftemperaturesensorsandtheimpactoftheserequirementsonthescaleofdeploymentandmeasurement.InSectionIIIwediscussourparadigmofRFIDtagantennabasedtemperaturesensing,howitwouldmeettheserequirementsanddescribetheconstructionofaprototypesensorbuilttoexaminethishypothesis.SectionIVthenattemptstoexaminethefeasibilityofthisprinciplebydrawingcomparisonstothebehaviourofadipoleantennainamulti-layereddielectricmedium.SectionVoutlinestheresultsofexperimentationonthesensorprototypeusedtovalidatethishypothesis.Finally,inSectionVIwesummarizethecontributionandoutlinethescopeforfuturework.
II.CURRENTSUPPLYCHAINTEMPERATURESENSINGALTERNATIVES
Temperaturemonitoringisparticularlyimportantinthecoldsupplychainwheretemperatureshavetobearti?ciallymaintainedtooptimizetheshelflifeofthetransportedgoods.Forexample,orangejuicetypicallyproducedandpackagedinFloridaismobilizedviatransportationnetworkstodifferentcornersoftheUS.Inordertoprolongthelifeofthejuice,thetransportationmediumisrequiredtobeappropriatelyrefrigerated.Similarlyvaccinesandotherperishablemedicalitemsmayrequiretobeconstantlystoredatsub-zerotemperaturesinordertoremainpotent.
Itiscrucialtologthecriticaltemperaturestatechangesduringtemperaturemonitoringofanenvironment.Mostofthecommercialsensors[V][VI][VII]availabletodayarecustomizedforpreciselythisfunctionalityandarecomprisedofaseriesofdiscreteelectroniccomponentssuchasantennas,applicationprocessors,batterypowerunits,signalconditioningunitsandmemory.Unfortunately,thepresenceofdiscreteelectroniccomponentsdrivesupthecostofthetemperaturesensorandthismanifestsinareductioninthenumberoffeasiblesamplepoints,duetomonitoringbudgetconstraints.Lackofadequatemonitoringgranularity,couldinadequatelyrepresenttheambienttemperaturepro?leandthereforeitisimportanttomaintainanappropriatebalancebetweensensorfunctionalityandcost.
Withthisinmind,wehighlightthemostimportantfactorsrequiredoftemperaturesensorsinsupplychainmonitoring:LowCost:MostcommercialtemperaturesensingoptionsaretypicallyavailableinthepricerangeofO($II0)today.Thepriceofthesesensorspreventtheirpervasivedeploymentinsupplychainapplications.Forexample,itdoesnotmakeanycommercialsensetotaganorangejuicecartonpricedatabout$IIIwithadedicatedsensorwhichcostsO($II0).Thustemperaturemeasurementswiththissensorislimitedtoinferencemeasurementsfroma?nitenumberofnodesdeployedinthefreezerorrefrigeratorstorageunitwhichcontainsthousandsofsuchcartons.Thegranularityoftemperaturemeasurementsmaynotbesuf?cienttocapture?uctuationsintemperatureacrossthefreezerunit.Inordertoattaintrueubiquityinsupplychainoperations,thesensorsshouldbeultra-lowcostsoastoimprovethescaleoftheirdeployment.
AbilitytoMaintainState:Mosttemperaturesensorstodaylogatimehistoryofdatasothatthresholdtemperatureexposureinformationcanberecorded.Thequestionofwhetheracriticalexposurethresholdwasattainedismoreimportantthanthetimehistoryitself.Itisthusimportantthatthesensorsbeabletologexposureinformationatminimumcost.StandardizationofCommunicationProtocols:Therearenumerouspartnersinteractinginasupplychain.Inordertofacilitateseamlessdatatransferfromonepartnertoanother,thetemperaturesensor’scommunicationprotocolmustbestandardized.Itisimportanttopicksensorsthatconformtoaprotocolthathastheleastsetupcostintermsofhardwareandsoftwareequipment.
RFIDpresentsawellestablished,standardizedwirelesstechnologytoleverageforsensordevelopment.Somecommercial
alternativeshaveattemptedtoprovideRFIDbasedtemperaturesensingsolutions.Forexample,KSW[VIII]hasdevelopedanRFIDsmartlabelwithanintegratedtemperaturesensor.Similarly,Gentag[IX]hasdevelopedanRFIDtagICwithanintegratedtemperaturesensor.BoththeseapproachesleveragetheRFIDGenIICommunicationprotocol[I.0]andthusconformtoastandardizedcommunicationprotocol.TheKSWsensorhowevercostsaround$III-$I.0andcannotbedeployedfortaggingcheaperitems,suchasperishablefoodproducts,whichwouldaccountforasizablefractionoftheitemspassingthroughthesupplychain.SimilarlythedesigncostsassociatedwiththeGentagICwoulddriveupthecostofthecompositesensortagwhenevertheICisintegratedwithatagantenna.ThereisthusaneedtodrivethecostofatemperaturesensorevenlowerinordertoensureitsubiquitousdeploymentandinthispaperweexaminethedesignofatemperaturesensorusingtheRFIDtagantennabasedsensingparadigm.TherehasbeensomepriorresearchworkinusingtheRFIDtagantennaasasensor.Forinstance,Marrocco[I.I.]presentbasictheoryandsimulationsfortheperformanceofmultiportRFIDtagsonmaterialswithdifferentpermittivities.Similarly,Bhattacharyyaet.al[I.II]presentatagantennabaseddisplacementsensorbyrelatingdisplacementtoachangeinreaderthresholdtransmittedpowertopoweruptheRFIDtagICandthedifferentialbackscatterpowerreturnedfromanRFIDtag,whileSidenet.al[I.III]illustratehowordinaryRFIDtagscanbeusedashumiditysensors.Inrelatedwork,Marroccoet.al[I.IV]presentexperimentalresultsfortagsensordesignsalongthelinesofthetheoryandsimulationspresentedin[I.I.].Inthenextsection,wediscussthedesignofanRFIDtagantennabasedtemperaturesensorthathastheadvantageofbeingultralowcostandlonglastingwithanabilitytomaintaincriticalthresholdtemperatureexposureinformationusingnon-electricmemory.
Fig.I..TemperatureSensorInitialization
Fig.II.TemperatureSensorWorking
III.TEMPERATURESENSOROPERATINGPRINCIPLE
RFIDtagantennabasedsensing[I.II]isaparadigmfordevelopingultralowcost,longlasting,standardizedandaccuratesensorsleveragingtheUHFRFIDinfrastructure.
ThissensingapproachreliesonmappingachangeinsomephysicalparameterofinteresttoacalibratedchangeinRFIDtagantennacharacteristicsandthusRFIDtagpowercharacteristics.
TherehasbeenpriorresearchintothedevelopmentofRFIDTagAntennabasedsensors.Forinstance,Bhattacharyyaet.al[I.V]discussthedevelopmentofadisplacementsensorusingthefacttheRFIDtagperformancedegradesincloseproximitytometal,whileSidenet.al[I.III]demonstratehownormalRFIDtagscanbeusedtodetectrelativehumidityandmoistureingress,bymakinguseofthefactthatRFIDtagperformancedegradesinproximitytowater.
Inthisstudy,wedesignanRFIDtagantennabasedtemperaturesensorforsupplychainscenarioswherethetemperatureofthetransportedgoodscannotexceed0oCforperiodsoftimeexceedingachosentolerancelimit.Ifthetemperatureismaintainedabove0oCformorethanthetolerance,thesensoriscapableofloggingthisandconveyingthisinformationtothenextRFIDreaderunitthatinterrogatesthetemperaturesensor.TheoperatingprincipleofthesensorissummarizedinFig.I.andFig.II.
ThetemperaturesensorisdesignedasacompactcuboidaljacketmadeoutofplexiglasswithtwoRFIDtaginlayspastedontheoutersurfaceoftheplasticjacket,separatedbyaboutIII-Vcm.Theinsideoftheboxis?lledwithanaqueousmediumhavingadesiredmeltingpoint-forthepurposesofthisstudy,wechoosewaterhavingameltingpointof0oC.Analuminiumplate,capableofverticalmotionintheaqueousmediumislocatedbehindthetwotagssuchthatatanygiveninstantoneofthetwotagsisdetunedduetotheproximityofthemetalplate.Thesensorreliesonchangesinphaseoftheaqueousmediumandverticalmotionofthealuminiumplateundertheforceofgravitytorecordchangesincriticalstateasweshallillustrate.
A.SensorOperatingPrinciple
Wenowanalyzeindetailtheoperatingprincipleofthissensor:
Initialization:Thetemperaturesensorisdesignedtomonitorthetemperatureofgoodsinafreezerbelow0oC.Thesensoristhusinitializedbyfreezingtheicewiththemetalplatedirectlybehindthetoptag-whichweshallhenceforthrefertoasTagAasseeninFig.I..Thusintheinitializedstate,wewouldexpectbetterperformancefromTagB,bothintermsofdifferentialbackscatterpowerfromTagBaswellasthresholdtransmittedpowerrequiredtopowerupTagB,sinceTagAisdetunedduetothemetalplatedirectlybehindit.Ifduringsupplychaintransportation,thegoodsremainatorbelow0oCforalltimeslessthanthetolerancelimit,themetalplateremainsfrozenbehindTagA.Eachreaderunitinterrogatingthesensorwouldbeabletoverifythatthegoodsweremaintainedattemperaturesbelow0oCthroughouttransit.Fig.I.illustratestheattachmentofthesensortooneoftheitemsintherefrigeratorpriortotransportation.
EffectofHighAmbientTemperature:Ifthetemperaturesensorisplacedattemperaturehigherthanthetemperatureofthefreezerforperiodsgreaterthanthetoleranceinterval,theicewouldmeltandthemetalplatewoulddescendunderthein?uenceofgravitytoapositionbehindTagBasseeninFig.II.Thefactthatthetemperaturesensorhasundergoneastatechangecanthenbedetectedbythefollowingobservations:
–ReductioninReadRange:Thewatermoleculeishighlypolarandtendstoorientitselfsoastocanceloutanyincomingelectric?eld,aphenomenonthatisnotaseffectiveinthesolidicestate.ThusaboutVIII0%oftheincomingelectric?eldwouldbere?ectedattheplexiglass-waterinterface[I.VI].Thisismanifestedinareductioninsensorreadrangefortagpowermeasurementsfrombothtags.Thusifasensorarrivesatadistributioncenterwiththeaqueousmediumintheliquidstate,thereadrangeforbothtagsisdrasticallyreduced.
–DetuningofTagB:SincethemetalplatedescendsundergravitytodetuneTagBinsteadofTagA,achangeofstatecanalsobecharacterizedbyadegradationofperformanceofTagBrelativetoTagA-whichisindependentofwhethertheaqueousmediumisintheliquidorsolidstate.Incontextofthesupplychain,ifacrateoffrozenproduce,taggedwiththissensor,iskeptoutinthesunforacoupleofhours,theiceinthesensorwouldmeltandthemetalplatewoulddescendtodetuneTagB.Evenifthegoodsareplacedbackinthefreezer,themetalplatecontinuestodetuneTagB.Thusitispossibletopermanentlyrecordthefactthatthetemperaturesensorsufferedacriticalstatechange,irrespectiveofwhatstatetheaqueousmediumassumessometimeafter.
Fig.III.TemperatureSensorinElevation
TheplanandelevationofthesensorisdescribedinFig.IIIandFig.IVrespectively.
B.PrototypeDesign
Weexaminethefeasibilityofthissensingapproachviathedevelopmentofaprototypesensor(c.fFig.V).Theplexiglassjacketwhichhousesthemetalplateandthe
aqueousmediumisdesignedforthedualpurposeofI.)testingtheperformanceofdifferentkindsofRFIDtaginlaysandII)allowingfordifferentvolumesofwatersothattheice-watertransitiontimecanbecalibratedandcontrolled.
SelectionofJacketSurfaceArea:RFIDtaginlaysvarysigni?cantlyindimensions,basedonantennageometryfromaboutI.0cmXIIcminthecaseofanAlienSquiggle[I.VII]toaboutIXcmXIXcminthecaseoftheImpinjBanjo[I.VIII].Thusthesurfaceareaoftheplasticjacketwassetasl=I.VIcmandb=I.0cm(Fig.III)tobeabletoaccommodatemostkindsofinlays.ThecompletelistofdimensionsisoutlinedinTable.I.ReferencesshouldbemadetoFig.IIIandFig.IVtointerpretthedimensions.
EffectofVolumeofWateronToleranceInterval:
InFig.VIweplotobservedtimerequiredfortheicetocompletelymeltasafunctionoffourdifferentcontrolvolumesvaryingfromIV0mltoI.00mlofwaterasshowninTable.IIforaconstantoutsidetemperatureofI.VIIIoC.We?ndthatalinearbest-?tfunctionadequatelycapturestherelationship.
FromFig.VIweobservethatitispossibletodrawafairlyaccuratecorrespondencebetweentheexperimentallyobservedmelttimeandacontrolvolumeofwater.Forthepurposesofexperimentation,acontrolvolumeofVVml,correspondingtoatoleranceintervalofaboutI.VIIImins,ischosen.Ingeneralhowever,thevolumeofwatercanbevariedsoastomeetthespeci?edtoleranceintervalbyreferringtoFig.VI.
Fig.IV.TemperatureSensorinPlan
Fig.V.TheSensorPrototype
TABLEI
DIMENSIONSOFTHEPROTOTYPE
Dimension)Value(mm)
JacketLength(l)I.VI0
JacketBreadth(b)I.00
JacketDepth(d)VI
JacketMaterialThickness(s)I..III
TABLEII
VARIATIONOFICEMELTTIMEWITHVOLUMEOFWATER
Volume(ml)MeltTime(min)
IV0I.VI
VIVII0
VIII0IIVII
I.00III0
Fig.VI.PlotofMeltTimeasafunctionofdifferentcontrolvolumes
IV.THEORETICALANALYSISOFSENSORPRINCIPLE
Changesinbackgrounddielectricmaterialwillcertainlyin?uenceRFIDtagperformancebothintermsofthepowerbackscatteredbythetagaswellasthethresholdtransmittedpowerrequiredtopowerupthetag.ThesensorprincipleoutlinedinSectionIIIreliesonthedifferenceinbackscatterpowerlevelsduetothepresenceorabsenceofthemetalplatebeingsigni?cantenoughtobeabletodifferentiateastatechange.Inthissection,weproposeaquickmathematicalanalysistodemonstratewhetherthisisindeedexpectedtobethecase.
Thenetfar?eldduetoadipoleantennaontopofadielectricsubstrateandassumingTransverseElectric(TE)radiationisduetothesuperpositionoftheelectric?elddirectlyradiatedfromthedipole,andthecomponentoftheelectric?eldre?ectedbythedielectricsubstrate,.Thissuperpositionisgivenby:
(I.)
whereistheeffectivere?ectioncoef?cientduetoasystemofdielectriclayersandisthefarelectric?eldofthedipoleantennawithnobackgrounddielectricinfreespace.
Fig.VII.DielectricLayersinthemodel
Weconsiderthreecaseshere.Inthe?rstcase,thebackgrounddielectrictotheRFIDtagistheplexiglassjacketfollowedbythemetalplate.Sincethemetalplateisassumedtobeperfectlyre?ecting,wedonotconsideranysubstratebeyondtheplate.Inthesecondcase,thebackgroundmaterialisplexiglass,followedbyalayerofice,followedbyplexiglassasshownbytakingsectionsatA-AandB-BinFig.IIIandillustratedinFig.VII.Thethirdcase,isthesameasthesecondcaseexceptthattheicedielectricisreplacedbywater.
AsoutlinedinKong[I.IX],thecoef?cientcanbecalculatedforannlayersystemby
(II)
Wede?netheterminologyusedinEq.IIbelow:
Heremistheeffectivere?ectioncoef?cientofthesystemofdielectricscorrespondingtothethreecasesm=I.,II,IIIformetal,iceandwaterdielectricrespectively.
relatestothethicknessofthelayerinthedielectricsystem.
isthezcomponentofthewavenumberinlayern.
isthere?ectioncoef?cientbetweentheandlayersandisgivenby:
(III)
herewehave
(IV)
whererelatestothemagneticpermeabilityoflayeriandisthecomponentofthewavenumberinmediumj.Fornormalincidence,relatestotheelectricalpermittivityforthelayerj.
TABLEIII
RETURNEDPOWERLEVELSOFICEANDWATERRELATIVETOMETAL
DielectricPowerRelativetoMetal
IceIIVIII.VVIIdB
WaterIII..I.VIIIdB
SincethePoyntingpowerdensityis,wehave
(V)
whereAistheconstantofproportionality.AssumingEisthere?ectioncoef?cientsforcaseI.,IIandIIIwehavethedifferenceinpowerlevelsoftheiceandwaterdielectric,relativetothemetalplatedielectricas
(VI)
and
(VII)
PlugginginthevariousvaluesfordibyreferringtoFig.VIIandmakingappropriateassumptionsfortherelativedielectricconstantsforwater,iceandmetalweobtainresultsofpowerdifferencelevelsassummarizedinTableIII.Wenotefromthesecomputedvaluesthatthedifferenceinpowerlevelsforiceandwateraresuf?cientlylargerelativetometalasabackgrounddielectric.Furthermore,thereisarelativedifferenceofaboutVIIdBbetweeniceandwaterandthusallcriticalstateandphasechangesareclearlydiscernible.Thusthefeasibilityandvalidityofthissensorshouldcertainlybeinvestigatedviaexperimentationandthisisthefocusofthenextsection.
V.EXPERIMENTATIONANDRESULTS
Inordertoverifythesensinghypothesis,wesubjectthesensortotemperaturecyclesfromthefrozenstatetothefullymeltedstate.AcontrolvolumeofVVmlofwaterisselectedforexperimentation.ReferringtoFig.VI,weexpectthemelttimetobeaboutI.IXminutes.
TagreadrangeisakeyissuetobedeterminedinthisexperimentsinceRFIDtagperformancedwindlessigni?cantlyinthepresenceofwaterandmetal.Thesensor-RFIDreaderseparationisvariedfrom0.VmtoI..Vminstepsof0.Vm.Foreachseparation,thesensoriscycledfromthefrozenstatetothemeltedstateandthereceivedbackscatterpowerfromTagsAandB(cf.Fig.III)asafunctionoftimeisplotted.WeexpecttheicetocompletelymeltinaboutII0minutes,howeverthetestisconductedforaI.hourinterval.
RFIDReaderSystem[I.VIII]isusedintheexperimentation.ThereadertransmittedpowersettingissettothemaximumofIVWEIRP.ThetemperaturesensorintheinitializedfrozenstateisplacedbeforethereaderantennaatanenvironmenttemperatureofI.VIIIoC.ThereaderantennameasuresthedifferentialbackscatterpowerfromTagAand
BasafunctionoftimeforaI.hourinterval.ReadingsaretakenattherateofIIpersecond,averagedoverVI0secondsandreportedonaperminutebasis.Wenowdiscussthesensorresponseasafunctionoftimeforthethreedifferentsensorreaderseparations.
Fig.VIII.BackscatterPowervs.timeforsensor-readerseparationof0.Vm
A.Reader-SensorSeparationof0.Vm
Fig.VIIIplotsbackscatterpowerresponsefromthetwotagsasafunctionoftime.Initially,TagAhasthemetalpiecebehindit,whichdetunesthetagantenna,asaresultofwhichTagAgivesalowerbackscatterpowerresponserelativetoTagB.TagBinitiallyhasiceasthebackgrounddielectric,whichdetunestheantennalessseverely,asaresultofwhichTagBinitiallyhasamuchbetterbackscatterpowerresponserelativetoTagA(cf.Fig.I.).
Fig.IX.BackscatterPowervs.timeaccountingforcondensationmoisturebuildup
Astimepasses,theicestartstomeltandthebackgrounddielectricforbothtagsinvolvesincreasingamountofwater.Watermoleculestendtoorientthemselvessoastocanceloutincomingelectric?eldsandthuswaterdetunesthetagantennaandre?ectsmostoftheincomingelectric?eld,thusaffectingperformancemoreseverelythanice.Thusbetween0andI.0minutes,theperformanceofbothtagsstartsreducing.BetweenI.0andIIIVminutesweseethatthereisnoobservedresponsefromTagA.Weobserveduringexperimentationthatwatervapourhasatendencytocondenseonthesurfaceofthecoldplexiglassjacketandarbitrarilydetunethetags.ThecondensedwaternowpresentsoneadditionallayerinfrontofthetagwhereVIII0%oftheincoming?eldisscattered.Itispossiblethatthiscondensationtogetherwiththeproximityofthemetalplate,detunesTagAtotheextentthatthetagICchipfailstopowerup.
Inordertoverifythishypothesis,werepeattheexperimentremovingmoistureduetocondensationassoonasitbuildsup.TheresultsareseeninFig.IX.Fromthe?gure,weobservethatTagAalwaysgivesaclearresponseoverIIIVminutes,illustratingthatmoisturecondensationmustbeaccountedforwhiletakingmeasurementsfromthesensorsystem.
ReferringbacktoFig.VIII,itisobservedthattheicecompletelymeltsandtheplatedescendstodetuneTagBataboutIIIVminuteswhichisreasonablyclosetothepredictedI.IXminutes.ThusafterIIIVminutes,TagBgivesaworseperformancethanTagA.Wenoteaninterestingobservationhere.AfterIIIVminutesTagAisnolongerdetunedbythemetal,butitsperformanceisworsethanthatofTagBatt=0,sincethebackgrounddielectricforTagAiswater,whichdetunesthetagantennamoreseverely,ratherthanice.Toverifythis,werefreezethesensorwiththemetalplatebehindTagB.AsseenfromFig.IX,theresponseofTagAthenimprovestolevelscomparablewithTagBatt=0.
Ifweattempttocorroboratetheexperimentalresultswithourtheoreticalanalysis,weobservethatthetrendspredictedbythetheoreticalanalysisarecon?rmedviaexperimentation.ThepowerlevelswithdielectricbackdropoficeisonanaverageaboutVIIdBmbetterthanwithabackdropofwaterasseenbycomparingthepowerlevelsofTagBatt=0withthoseofTagAbeyondt=III0inFig.VIII.Similarly,thereisasigni?cantdifferenceofaboutI.0dBmbetweenabackdropofmetalvisa-visice.AlthoughnotashighasthetheoreticallycomputedIIVIIIdBm,theassumptionoftherebeingasigni?cantdifferenceinpowerlevelbetweenthestatesiscertainlyveri?edinpractice.
Inordertovalidatetherepeatabilityoftheobservations,thetemperaturecycletestwasrepeatedthreetimesandineachrun,repeatableresultswereobtained.
B.Reader-SensorSeparationofI.andI..Vm
Fig.VIIIshowstheresponseofthesensorasafunctionoftimewhentheseparationwasincreasedtoI.m.Thetrendsofthecurvesremainthesamealthoughtheabsolutevaluesofthepowerlevelsdecreaseslightlyduetotheincreaseddistancebetweenreaderandtag.ItisinterestingtonotethatatI.mdistances,oncethemetalplatedescends,TagBdoesnotgetpickedupbythereader.Thepowerdensityreachingthetagovertheincreaseddistanceaftermismatchlossesduetothewaterandthemetalisinsuf?cienttopowerupTagB’sIC.Similartrendswereobservedwhenthereader-tagseparationwasincreasedtoI..VmasshowninFig.I.I..Asweobservefromthe?gure,thereareseveralmissedobservations,indicatingthatthebackscatterpowerreadingsfromthesensoraredif?culttodetectatthatrange.
Fig.I.0.BackscatterPowervs.timeforsensor-readerseparationofI.m
Fig.I.I..BackscatterPowervs.timeforsensor-readerseparationofI..Vm
Asinthepreviouscase,thesetestswererunthreetimesaswellandconsistentresultswereobtained.Fig.I.IIforinstancedemonstratesthreerunswhenthesensor-readerseparationisI.m.
Fig.I.II.ThreeRunsforReader-SensorSeparationofI.m
Fig.I.III.Effectofdifferentmaterialofdeployment
C.EffectofMaterialofDeployment
Itisalsoimportanttoconsidertheelectromagneticin?uenceofthematerialonwhichthesensorisdeployed.Theexperimentswereconductedbyattachingthesensortoahollowcardboardbox,howeveritisnecessarytoconsidertheeffectofothermaterials-particularlythesocalledRFunfriendlymaterials.Alotofpackingcratesandcontainersaremetallicandinordertosimulatethisthesensorisattachedtoametallicplateandtheexperimentisrepeatedfora0.Vmsensor-readerseparation.TheresultsareshowninFig.I.III.Aswecanseethepresenceofmetalsigni?cantlyalterstherelativepowerdifferencesbetweenthetwotags.Infact,themetalseemstoboostthestrengthofthebackscatterpowersignal.WenotethatthemetalplateislocatedatadistanceofaboutIIcmfromthetags.Itispossiblethatthebackscatterpowerfromthetagandthere?ectedsignalfromtheplateareinterferingconstructivelyatthereaderandthismayaccountfortheboost
insignal.Thusmaterialofdeploymentisanimportantfactortobeconsideredandwillbeaddressedinfutureworktoenhancethesensorperformance.
VI.CONCLUSIONS
Inthispaper,weproposeadesignmethodologyandverifytheworkingfeasibilityofanultralowcostwirelesstemperaturesensorwhichleveragestheRFIDinfrastructure.Whilemanyofthesensorscommerciallyavailableinthecoldsupplychain,boasthighprecisionandaccuracyofmeasurementwithfeaturessuchasmemorylogging,theirrelativelyhighcostprecludestheirpervasivedeploymentincoldsupplychainapplications.
Ourapproachemphasizestheultra-lowcostofmanufacturebutrecognizestheassociatedtrade-offsthiswouldentail.Tagantennabasedtemperaturesensingwillnotproviderealtimeupdatesorlogatimehistoryoftemperaturemeasurements.Howeverthissensorwillprovidetheuserwithbinarystateinformationindicatingwhethercriticaltemperaturethresholdswerereachedduringtransportationallowingtheusertomakeadecisionofwhethertokeepordiscardthegoods.Sincethestateofmostgoodsareexaminedatthedistributioncentersinasupplychain,ratherthanintransit,wearguethatthiskindofsensingmightinfactbeadequateformonitoringmostitemssuchasorangejuicecartons,milkandmeatproductsinsupplychainapplications,sincethestateoftheproductswillbeautomaticallymonitoredassoonastheproductspasspastanRFIDreaderatthedockdoorofadistributioncenter.Furthermore,sincemanysupplychains
haveanRFIDinfrastructurealreadyinplace,thissensingparadigmhastheadvantageofleveraginganalreadywellestablishedcommunicationinfrastructure.Thuswithitsultra
lowcostandabilitytoconveycriticalstatechangeinformationuponinterrogation,weexpectRFIDtagantennabasedsensingtobeveryrelevanttoalargemarketshareofcoldsupplychainapplications.
Whilethedesignmethodologywasdemonstratedinthispaper,thereissigni?cantresearchthatrequirestobedoneinthisareatoextendthetechnology.WenotethatthelowerTagBisconsistentlynotreadforreader-tagseparationsofmorethan0.Vm.Itisdif?culttodifferentiatethisfromafailureofTagB.ArelativeresponsefrombothtagsAandBwouldbeamuchbetterindicatorofstatechange.Itisthusnecessarytodesignatagwithcustomantennageometrythatisoptimizedforperformancenearwaterandmetalsothatthereadrangeofthesensorisboostedtomorethan0.Vm.
ThecurrentprototypeimplicitlyassumesthatthesensorwouldbedeployedonpackagingmaterialthatisRFfriendly.Inpractice,manytransportationpalletsaremetallicandtheperformanceofthesensorisaffectedbythis.Modifyingthesensordesign,perhapsbyincludingametalbackplane,wouldmaketheperformanceindependentofthematerialofdeployment.
Condensationofmoistureonthesurfaceofthesensorisanotherissuethatmustbetackled.Evenforreader-sensorseparationsofabout0.Vm,thebuildupofmoisturecandetunethetagandre?ectincidentelectric?eldstotheextentthatthetagisnolongerpickedup.Eliminationofmoisturebuildupbyuseofappropriatehydrophobicsurfacesthatshedwateroncontactisanotheravenuethatmustbeexploredinthesensordesign.
Therequirementthatthissensorbeinitializedbyfreezingthemetalplateinpositionbehindthetoptagisonepotentialimpedimenttothecommercialdeploymentofthissensor.Althoughitisnotdif?cultforsupplychainpartnerstoinitializebatchesofthesesensorsinindustrialfreezers,itisoneadditionalstepthatdetractsfromthelow-cost,pervasiveappealofthissensor.Furthermore,sincethememoryeffectistriggeredbygravity,itisnecessarytoplacethesensoronthesideoftheshippingcontainer.ThusanunderlyingassumptionisthatthegoodsarenotrotatedorturnedbyI.VIII0degreesduringtransit.Furthermore,thiscurrentversionofthetemperaturesensorisnottamperproofandthememorymechanismbreaksdownshouldthesensorberotatedbyI.VIII0degreesandrefrozen.Ourresearcheffortsarecurrentlylookingintodevelopingalow-costsolidstatesensorthatisindependentofsensorplacementorientationandwhichdoesnotrequiretobeinitializedbeforedeployment.
Extendingthesensortodetecttheviolationofseveralthresholdsisanotherimportantstepperhapsviatheuseofcompositedielectricshavingdifferentmeltingpoints.Itmightbeusefultodesignasensorthatdetectsnotonlyaupperboundcritical
temperaturebutalowerboundaswell.Forexamplemostorganicproducewouldbeoptimizedfortransportfortemperaturesbetween0andI.0C.ThefoodwouldspoilifmaintainedforalongtimeaboveI.0Cbutwouldalsospoiliffrozenbelow0C.DevelopinganRFIDbasedtemperaturesensorhavinganoptimizedrange,withextensionsformonitoringdifferenttemperaturerangethresholdsandwhichisindependentofthepackagingmaterialofdeploymentwouldthusbeasigni?cantboosttotheproposedtechnologyandisthusthefocusoffutureresearch.
REFERENCES
[I.]M.CO’Connor,RFIDJournalArticle:Coldchainprojectrevealstemperatureinconsistencies,"
http://www.r?djournal.com/article/view/IIVIIIVI0/I.,DecemberI.,II00VI.
[II]S.Estrada-Flores,Acousticmeasurementofavocado?rmnessduring
lowtemperaturestorage:Effectofstoragetemperature,"International
ConferenceonQualityinChains.AnIntegratedViewonFruitand
VegetableQuality,pp.IIIIIIII–IIIIIIX,II00III.
[III]E.WSchuster,S.JAllenandD.LBrock,GlobalRFID:TheValueofthe
EPCGlobalNetworkforSupplyChainManagement,"SpringerVerlag,
II00VII.
[IV]A.DadaandF.Thiesse,SensorApplicationsintheSupplyChain:The
ExampleofQuality-BasedIssuingofPerishables,"LectureNotesin
ComputerScience,vol.IVIXVII/II00VIII,pp.I.IV0–I.VIV,II00VIII.
[V]EvidenciaTemperatureSensingSolutions,ThermAssureRFTemperatureSensor,"http://www.evidencia.biz/products/acti-tag.htm.
[VI]PultronicsWirelessDesignSolutions,TechnicalLiteratureonWarehouseTemperatureMonitoringSolution,"
http://www.pultronics.com/warehousetemperature.htm.
[VII]C.S.T.Solutions,Coldtrakcoldmonitoringsolutions,"
http://www.supatrak.com/cold-chain-monitoring.html.
[VIII]RFIDNewsArchives,NewLost-CostTemperatureSensor,"
http://www.r?djournal.com/article/view/IIVIII.
[IX]UsingRFIDNewsArchives,Low-CostTemperatureSensortoaddtoRFIDInlays,"
http://www.usingr?d.com/news/read.asp?lc=wIIIIIVVrxVIIVVzo.
[I.0]TheEPCGlobalStandards,ClassI.GenerationIIUHFAirInterface
ProtocolStandard,"http://www.epcglobalinc.org/standards/uhfcI.gII/,
II00VIII.
[I.I.]G.Marrocco,L.MattioniandC.Calabrese,MultiportSensorRFIDs
forWirelessPassiveSensingofObjectsBasicTheoryandEarlyResults,"
IEEETransactionsonAntennasandPropagation,vol.VVINo.VIII,pp.
IIVIIXI.–IIVII0II,II00VIII.
[I.II]C.Floerkemeier,R.BhattacharyyaandS.Sarma,BeyondtheIDin
RFID,"II0thTyrrhenianInternationalWorkshoponDigitalCommunications-aCNITConference,II00IX.
[I.III]Johan,S.andXuezhiZengandUnander,T.andKoptyug,A.and
Nilsson,H.-E.,RemoteMoistureSensingutilizingOrdinaryRFID
Tags,"Sensors,II00VIIIEEE,pp.III0VIII–IIII.I.,Oct.II00VII.
[I.IV]G.MarroccoandF.Amato,Self-sensingpassiver?d:Fromtheoryto
tagdesignandexperimentation,"MicrowaveConference,II00IX.EuMC
II00IX.European,pp.00I.–00IV,IIIXII00IX-oct.I.II00IX.
[I.V]R.Bhattacharyya,C.FloerkemeierandS.Sarma,TowardsTagAntenna
BasedSensing-AnRFIDDisplacementSensor,"IEEERFIDII00IX
ConferenceProceedings,pp.IXV–I.0II,II00IX.
[I.VI]D.M.Dobkin,TheRFinRFID:PassiveUHFRFIDInPractice,"
NewnesPublishing,II00VII.
[I.VII]AlienTechnologies,AlienTechnologyProducts,"
http://www.alientechnology.com/products/index.php,II00IX.[Online].
Available:http://www.alientechnology.com/
[I.VIII]ImpinjRFIDProducts,ImpinjUHFGenIIRFIDProducts,"http://www.impinj.com/products/default.aspx.[Online].Available:
www.impinj.com
[I.IX]J.AKong,ElectromagneticWaveTheory,EMW,Ed.JohnWileyand
Sons,I.IXVIIIVII.
RFID标签天线进行温度测量
RahulBhattacharyya,ChristianFloerkemeierandSanjaySarma
汽车身份证中心
麻省理工
剑桥,麻萨诸塞州0III.IIIIX
电子邮件:rahulb,floerkemandsesarma@mit.edu
摘要－温度监测在I.些领域中是很重要,尤其是在寒冷供应链的应用.大多数商业无线温度传感器包括收发器,存储器,电池组和电池保持温度长久,这些不但很昂贵,并且部署的传感器有限.在本文中,我们提出了I.个无源RFID标签天线的低成本温度传感器.使用I.个简单的机械方法,永久诱导RFID标签暴露在大于阈值时的温度的功率特性变化.当违反临界温度阈值时读者就可以检测或监测接收到的回波信号强度.通过理论和实验手段的研究验证假设的可行性.这将表明,这种检测方式的潜力有利于提高温度传感节点的普遍性和提高供应链的可视性和性能.
I..介绍
温度是I.个重要的环境参数,在很多领域都有应用,如基础设施监控领域.化学.环境工程和冷供应链管理.综上所述可知在寒冷的供应链运作中奥康纳温度监测就显得尤为重要[I.],它指出I.条运输托盘内的温度变化高达IIIV％.正如埃斯特拉达-弗洛雷斯[II]和舒斯特[III]等人所述,环境温度波动会影响易腐产品的质量.在供应链运作中估计III0％的易腐货物[IV]在运输过程中会损坏,增加检控的传感设备的精确度有着迫切需要.
温度传感器在监测和记录环境临界温度是必要的.商业供应链应用中的无线温度传感器种类繁多,包括通过适当的车载电子实现的传感器,如内存和电池驱动起来的成本传感器单元.因此,I.个真正大规模部署的领域在考虑经济情况时往往会排除这些传感器.通过有限的传感器节点检测来推导出坏境温度 *好棒文|www.hbsrm.com +Q: ￥3^5`1^9`1^6^0`7^2$
的变化,而不是由通过供给链监视各物流单元.因此,这将是非常有用的,有I.个专门的在物流单元液位传感器使关键温度状态的变化可以被监测和记录.然而,为了达到这个目标,该传感器的费用必须显著低.在本文中,我们提出了无源UHFRFID的原理的超低成本的温度传感器,它能够在上述临界温度阈或用户规定的公差区间内记录温度变化.
第II节讨论了对温度传感器的要求以及这些要求对部署和测量尺度的影响.在第III节中,我们讨论了RFID标签天线的温度传感的典型范例,研究它将如何满足这些要求的和描述的I.个结构传感器的原型.然后第IV节探讨这I.原则的可行性,以及在多层电介质中偶极天线.第V节讲述了实验中所用的传感器原型,验证这I.假说的正确性.最后,在第VI部分我们概述了对于未来的贡献和大规模的推广.
II.如今在供应链温度测知中的选择
特别是在寒冷的供应链中,温度必须要人为地保持,从而优化运输货物的货架寿命.例如,橙汁在生产和包装的过程中,通过运输网络从佛罗里达州送到美国不同的的角落.为了延长果汁的使用寿命,需要在运输过程中适当冷藏.同样疫苗和其他易腐烂的医疗项目的也需要长时间储存在零度以下.
环境温度监测和记录的关键是在临界温度状态变化期间.今日大部分可定制的商业传感器[V][VI][VII]都具有这个功能,并且是由I.系列的离散的电子元件组成,如天线,应用处理器,电池动力单元,信号调理单元和内存.不幸的是,分立电子元件的存在抬高了温度传感器的成本,由于监控预算约束,则在误差范围内减少采样点.因此缺乏适当的监测精度.想要充分描绘环境温度曲线,重要的是在传感器的功能和成本之间取得适当的平衡.
考虑到这I.点,我们强调下供应链监控对温度传感器的要求:低成本,今天大多数商业级温度传感通常在O(II0美元)价格范围.这些传感器的价格使它们在供应链中不能普遍应用.它不是任何商业意义上的售价约III美元,它有专门的O(II0美元)传感器的成本.因此,温度测量传感器只限于从部署在冷冻或冷藏存储单元节点的有限数量的测量得出推论,其中包含很多这样的传感器.温度测量的精度可能不足以捕获整个冷冻机组的温度波动.为了实现真正的无处不在的供应连锁经营,传感器应该是超低成本,以提高他们的部署规模.
能否维持状态:今天大部分的温度传感器记录数据,使阈值温度曝光信息可以被记录下来.现在I.个关键的问题是门槛曝光是否达到比时间历史更重要.该传感器能够以最低的成本记录曝光信息是很重要的.
通信协议的标准化:有许多合作伙伴供应链中交互.为了便于从I.个伙伴数据无缝传输到另I.个时,温度传感器的通信协议必须被标准化.挑选符合最少成本的协议是非常重要的.
I.个RFID的.完善的.规范化的无线技术促进了传感器的发展.I.些商业已经尝试了RFID温度传感替代解决方案.例如,KSW[VIII]开发I.个带有集成温度传感器的RFID智能标签.同样,Gentag[IX]开发了I.种RFID标签芯片与集成温度传感器.这两种方法用的是RFIDGenII的通信协议[I.0],符合标准化的通信协议.然而,KSW传感器的成本约为III0-I.0元,不能被部署应用于标记更便宜的物品,如易腐食品的产品,这将占到供应链的项目相当大的比例.同样的只要该IC集成了I.个标签天线,GentagIC相关的设计成本会抬高复合传感器标签的成本.
因此,为了保证其大规 *好棒文|www.hbsrm.com +Q: ￥3^5`1^9`1^6^0`7^2$
模的应用,有必要继续降低温度传感器的成本.在本文中,我们探讨了I.个RFID标签天线传感模式的温度传感器的设计.例如,目前Marrocco[I.I.]的基本理论和模拟,适用于材料多端口RFID标签具有不同的介电常数.同样,巴塔查等人[I.II]提出了I.个标签天线,传感器发射功率电的RFID标签IC和从RFID标签反向散射,而Siden等人[I.III]解释了普通的RFID标签是如何用作湿度传感器的.在相关工作中,Marrocco[I.IV]目前的实验结果表明标签传感器设计是沿在理论中模拟出来的[I.I.].在下I.I.节中,我们讨论了温度传感器的RFID标签天线的设计,它具有超低成本.维护使用非电记忆临界阈值温度曝光信息的能力和持久的优点.
图I.温度感应器设定初值
图II温度感应器工作
III.温度感应器操作原则
RFID标签天线传感[I.II]是I.个典型例子,利用超高频RFID开发超低成本.持久化.规范化.精确的传感器.该传感方法依赖于RFID标签天线的特性,因此RFID标签具有电源特性.
目前已经在对RFID标签天线的传感器的研究有所成果的也有不少.例如,巴氏等[I.V]用事实验证了位移传感器中RFID标签性能下降密切接近金属,而Siden等人[I.III]演示了如何正常使用RFID标签检测相对湿度和湿气侵入,通过实验使该RFID标签性能降低接近水.
在这项研究中,我们设计了RFID标签天线温度传感器,用于在供应链中选择运输货物的温度在不超过0℃的时间段的I.个容忍限度.如果将温度保持在高于0℃以上,上述传感器能够记录当前和输送信息到下I.个RFID读取器单元.传感器的工作原理总结在图I.和图II.
上述温度传感器被设计成I.个紧凑的由立方体外套做出来的有机玻璃,两个RFID标签嵌体粘贴在塑料护套的外表面,隔约III-V厘米.箱的内部填充有水介质,为了本次研究,我们选择有水的熔点0摄氏度的铝板,能够垂直在水性介质中运动时在两个背后标记,使得在任何给定时刻时两个标记中的有I.个是失谐的.该传感器依靠改变在水性介质中的相位和力的作用,利用该铝板的垂直运动重力来记录变化的临界状态.此外,我们将举例说明.
I..感应器操作原则
我们现在详细地分析这感应器的操作原则:
初始化:设计I.个监视货物在0℃以下冷冻温度的传感器.现在开始我们称之为标签,如图I.所示,该传感器是冰正后方顶部的标签,与金属板冻结从而完成初始化.因此,在初始化状态,我们希望有更好的性能.由于金属板后面直接传送所需的阈值功率,无论是在从标签B差分后向散射功率,以及上电标记B,标签B都是失谐的.如果在供应链运输中货物保持在或低于0°C的时间小于公差极限时,金属板将仍然冻结背后的标签A.每个读取单元询问传感器确保货物在整个运输途中是温度保持在0℃以下.图I.所示的是运输前冷藏库中的I.个项目传感器的附件.
高环境温度的效果:如果在高于冷冻室的温度期间,温度传感器所测的温度大于温度公差区间,冰会融化,金属板将重心向着标签B后面的位置,如图II所示,温度传感器发生了状态变化,我们通过以下几方面来观察:
(I.)减少读取范围:该水分子是极性高,而且往往自我定向,从而抵消任何传入的电场,这种现象是不I.样的有效固体冰状态.因此大约VIII0％的输入电场将反映在有机玻璃-水界面[I.VI].这表明传感器读取范围的减少.因此,如果I.个传感器到达与水性介质中的分配中心在液体状态下,读出范围为标签大幅度减少.
(II)失谐标签B:由于金属板在重力作用下下降到失谐标签B,而不是标签的状态的变化也可以通过表征相对标签B的性能降低到标签的,与水的介质是否为液体或固体状态是不相关的.在供应链中,如果I.箱带有该标志的传感器冷冻农产品,为了在阳光下保持几个小时,在传感器中的冰会融化.金属板会沦落到连接失谐标记B.如果货物被放回冷冻室中,金属板将继续连接失谐标签B.因此,根据某个时间记录,假定无论处于何种状态的水性介质中,温度传感器将遭受严重状态变化.
传感器的平面图和正视在图III和图IV分别说明.
图III温度检测在升高
II.原型设计
让我们看看应用这种检测方法的传感器发展的可行性(参见图V).
有机玻璃护套容纳的金属板和水性介质被设计为双用途I.)检测不同种类的RFID标签嵌体的性能II)使用不同量的水,使得冰/水的过渡时间可以进行校准和控制.
护套表面区域的选择:RFID标签的尺寸有明显的不同,在外来曲线[I.VII]的情况下,天线的几何尺寸从约I.0厘米×II厘米,而在Impinj班卓琴[I.VIII]的情况下约IX厘米×IX厘米.因而塑料护套的表面区域被设置为I=I.VI厘米,B=I.0厘米(图III),以便能够容纳大多数种类的嵌体.尺寸的完整列表如表I.概述.参考文献中概述应用图III和图IV来解释尺寸.
图IV温度检测的方案
图V感应器原型
水的量对公差区间的效果:在图VI中,我们绘制所需的冰完全融化为IV个不同的控制量从IV0毫升变到I.00毫升水,如表功能要求.II为I.个恒定的温度I.VIII℃.我们发现,I.个线性的最佳拟合函数充分抓住了关系.
表I.原型的尺寸
尺寸价值(毫米)
护套长度(l)I.VI0
护套宽度(b)I.00
护套深度(d)VI
护套材料的厚度(s)I..III
表II冰的融化时间的变化及用水的体积
体积(毫升)融化时间(最小)
IV0I.VI
VIVII0
VIII0IIVII
I.00III0
图VI熔融时间为积的不同控制体积的函数
从图VI我们看到,控制时间与水体积之间有着相当准确的对应关系.实验的内容是控制I.个VV毫升的体积对应于I.个大约I.VIII分钟的容忍区间的选择.但总体来说,水的体积可以变化,以参照图VI满足规定的公差区间.
IV.感应器原则的理论上分析
介质材料的变化肯定会影响RFID标签的性能,无论是在动力方面的后向散射的标签,还是阈值所需的功率传输功率的标签.在第III节中概述的传感器原理依赖于在反向散射功率水平的差异是由于金属板是足够显著能够区分的状态变化的存在或不存在.在本节中,我们提出了I.个快速的数学方法来证明.
由于在电介质基板的偶极子天线和假设横电()净远场辐射是由于电场的叠加从偶极直接辐射,和电场的介质基片的反射,.这种叠加了:
(I.)
其中是有效的反射系数,是电介质层的I.个系统和是远电场偶极天线在自由空间中无背景电介质.
图VII电介体在模型中分层堆积
在这里我们考虑III种情况.在第I.种情况下,在背景电介质的RFID标签是有机玻璃护套随后将金属板.由于金属板被假定为完全反射的,我们并不认为超出了板的任何基底.在第II种情况下,在背景材料是有机玻璃,其次是I.层冰,随后有机玻璃所示,采取在各节A-A和B-B在图III和图VII中示出.第III种情况,是相同的,不同的是冰电介质被替换为水的第II壳体.
由于在Kong[I.IX]所述,系数RTE可以计算出I.个n层系统
(II)
我们定义在方程II中所使用的术语:
这里m为对应于分别m=I.,II,III为金属,冰和水的介电常数的III种情况的电介质系统的有效反射系数.
涉及的电介质系统的第i层的厚度.
是在n层波数的z分量.
表示是第i个和第j个层之间的反射系数,并且由下式给出:
(III)
在这里,我们有
(IV)
其中导率涉及层的磁导率和是在介质j中的波数分量.对于垂直入射,涉及介电常数中的第j层.
表III返回的冰和水的相对功率水平的金属
电介质功率相对应的金属
冰IIVIII.VVIIdB
水III..I.VIIIdB
由于坡印廷功率密度为,我们有
(V)
其中A为比例常数.假定是对于壳体I.,II和III,我们在冰和水介质的功率电平的差的反射系数,相对于所述金属板电介质作为
(VI)
和
(VII)
参考图VII并提出相对介电常数为水.冰和金属得到的功率差水平的结果,如表III,适当的假设,堵在各个值II.我们注意到从这些计算值,在功率水平的冰和水的区别是足够大的相对于金属作为背景介质.此外,还有I.个相对的约VII分贝冰和水,因此所有的临界状态和相位的变化之间的差异是明显可见的.因此,该传感器的可行性和有效性当然应该通过实验研究,这是下I.节的重点.
V.实验研究与结果
为了验证传感假说,我们受到传感器温度周期从冻结状态到完全熔融状态.选择VV毫升水的控制量为实验.参照图VI,我们预计熔体的时间约为I.IX分钟.
在这个实验中要确定标签读取范围,因为RFID标签的性能,在水和金属的存在显著不断下降的I.个关键问题.传感器-RFID读取器分离从0.V米变化到I..V米的步长为0.V微米.对于每个分离,传感器从冻结状态循环到熔融状态,并从标签A和B所接收的反向散射功率(参照图III)作为时间的函数作图.我们预计冰完全融化在II0分钟左右,但是测试进行了间隔I.小时.
实验中使用的是Impinj的RFID读取器系统[I.VIII].所发送的读取功率设定为IVWEIRP的最大值.在初始化的冻结状态的温度传感器被放置在读取器天线在I.VIII℃的环境温度之前.阅读器天线测量从标记A和差分散射力B为时间为I.小时的时间间隔的函数.读数采取以每秒II张的速度,平均超过VI0秒,并报告了每分钟计算.我们现在讨论的传感器的响应随时间的III个不同的传感器-读取器的分离的功能.
I..0.V米读卡器,传感器分离
图VIII绘出后向散射功率响应来自两个标记为时间的函数.最初,标记A具有金属片的背后,这走调的标签天线,其结果的标签A给出I.个较低的散射功率的响应相对于标签B.标签乙最初具有冰作为背景介质,其失谐天线少严重,因此其标签B最初有I.个更好的后向散射功率响应相对于标签的(参见图I.).
图VIII后向散射功率与0.VM传感器器分离时间
随着时间的推移,冰开始融化和背景电介质两个标签涉及的水增加量.水分子倾向于自己定位,以便抵消收到的电场,因而水走调的标签天线和反射大部分进入的电场,从而比冰更严重影响性能.从而在0和I.0分钟之间,这两个标签的性能开始降低.I.0和IIIV分钟之间,我们看到,有从标签A则完全没有观察到的反应,我们的实验中观察到,水蒸汽有I.种倾向,凝结的表面上冷有机玻璃外套和任意失谐的标签.冷凝水现在呈现在其中的VIII0％的输入字段的被散射的标记的前I.个附加层.这是可能的,这种缩合连同金属板附近,走调标记A至该标签的IC芯片不通电的程度.
为了验证这I.假设,我们重复实验去除湿气由于凝结尽快建立起来.其结果见于图IX.从图中,我们观察到标签的总是给人在IIIV分钟内I.个明确的答复,说明湿气凝结,必须考虑同时进行测量的传感器系统.
返回参照图VIII,它是观察到的冰完全融化和板下降到失谐标签乙在约IIIV分钟这是相当接近的预测I.IX分钟.因此,IIIV分钟,标记B给出了I.个糟糕的性能比标签答:我们注意到I.个有趣的现象在这里.第IIIV分钟标签的是由金属不再失谐,但它的性能比在t=0标签B的差,因为背景介质为标签的是水,而走调标签天线更严重,而不是冰.为了验证这I.点,我们重新冻结的传感器标签B.后面的金属板从图IX看出,标签A的响应,那么提高到与标签B在t=0相媲美的水平.
图IX后向散射功率与占冷凝受潮时间
如果我们试图证实我们的理论分析实验结果,我们观察到的理论分析预测的趋势通过实验得到证实.功率电平与冰的介电背景是在I.个平均约VIIdBm的比水的大背景下更好,因为看到在t=0与超越T=III0图VIII标签A的标记相比较B的功率电平.同样,有I.个约I.0dBm的金属签证-可见冰的背景之间有显著差异.?虽然没有那么高的理论计算为IIVIIIdBm,有可能出现的状态之间的功率电平的显著差异的假设肯定是验证在实践中.
为了验证该意见的可重复性,在温度循环试验重复III次,并且在每次运行中,得到可重复的结果.
II.I.至I..V米读卡器,传感器分离
图VIII示出了传感器的响应为当分离增加到I.m的时间的函数.该曲线的趋势不变,尽管电源电平的绝对值,由于阅读器和标签之间的距离增加稍有降低.有趣的是注意到,在I.米的距离,I.旦金属板下降时,标签B没有得到拾起的读者.的功率密度达到超过由于水失配损耗后增加的距离和所述金属标记不足以电标签B的集成电路.类似的趋势,观察当读者标签分离增加至I..V米,如图I.I.所示.正如我们从图中观察,有几个未接的观察,表明来自传感器的背散射功率读数是难以检测在该范围内.
图I.0后向散射功率与I.米的传感器读取器分离的时间
图I.I.后向散射功率与I..V米的传感器读取器分离的时间
图I.II为I.m读传感器分离III次
图I.III部署不同材质的影响
如在前面的情况下,这些试验进行III次,以及和I.致的结果得到.图I.II为实例演示了III个运行时,传感器读取器分离为I.m.
III.材料的配置影响
同样重要的是要考虑在其上的传感器被部署的材料的电磁影响.本实验是通过将传感器连接到中空板纸箱进行的,然而,有必要考虑其它材料的影响-尤其是所谓的RF不友好的材料.很多包装包装箱及容器是金属的,以模拟该传感器被连接到I.个金属板和重复实验为0.Vμm的传感器-读取器的分离.其结果示于图I.III.正如我们可以看到金属的存在显著改变了两个标记之间的相对功率差异.事实上,金属似乎刺激后向散射功率的信号的强度.我们注意到该金属板位于从标签约II厘米的距离.它可能是由从板的标签和反射信号的后向散射功率的建设性干扰在读取器和这也许可以解释在信号升压.从而调配材料应被认为与将在以后的工作加以处理,以提高传感器的性能的重要因素.
VI.结论
在本文中,我们提出了I.种设计方法,并验证了超低成本的无线温度传感器,它充分利用了RFID基础设施的应用的可行性.虽然在许多寒冷的供应链市售传感器中,拥有高精度和测量精度的功能,如内存记录,他们的成本相对较高,妨碍他们的普遍部署在寒冷的供应链应用.
我们的做法强调了制造成本超低,但承认相关的权衡这需要.标签天线温度传感将不提供实时更新或记录温度测量的时间历程.然而这种传感器将提供具有指示在运输过程中是否达到临界温度阈值,允许用户作出的是保留还是丢弃该货物的决定II进制状态信息的用户.由于大多数货品的状态进行检查在供应链,而不是在运输途中的配送中心,我们认为,这种传感实际上可能是适当的监测大部分项目,如橙汁纸箱,牛奶和肉类产品供应链应用中,由于产品的状态将被自动只要产品通过过去的RFID读写器在I.个配送中心的闸门监控.此外,由于许多供应链有I.个RFID基础设施已经到位,该传感模式具有充分利用已经建立的良好的通信基础设施的优势.因此,其超低的成本和传达临界状态变化信息的能力经审讯,我们预计RFID标签天线传感是非常相关的冷供应链应用了大量的市场份额.
虽然设计方法被证明在本文中,有需要应在这方面做延长技术显著研究.我们注意到,较低的标签B始终没有阅读超过读写器标签分离0.Vμm以下.这是很难区分这从标签B的故障来自标签A和B的相对反应是状态变化的I.个更好的指标.因此,有必要设计具有定制的天线几何形状为接近水和金属的性能优化,使得在传感器的读出范围被提高到大于0.V微米的标记.
当前原型隐含地假设,该传感器将被部署于包装材料,是射频友好.在实践中,很多运输托盘是金属和传感器的性能受到影响.修改所述传感器的设计,也许包括I.个金属背板,将使性能独立部署的材料制成.
水分传感器的表面上凝结是另I.个问题,必须加以解决.即使对于0.Vm左右读者传感器分离,水分的积聚可失谐标签和反射入射电场的范围内,标签不再回升.消除积聚的水分通过使用该洒在接触水适当的疏水表面是另I.种途径,必须在传感器的设计进行探讨.
该传感器通过在顶部标签后面位置冻结的金属板进行初始化的要求是I.个潜在的阻碍该传感器的商业部署.虽然这不是困难的供应链合作伙伴来初始化这些传感器的批次在工业冷冻机,它是有损于该传感器的低成本,普适上诉I.个额外的步骤.另外,由于记忆效应是由重力引发的,有必要将传感器在包装盒的I.侧.因此,I.个基本的假设是,货物在运输过程中不旋转或旋转I.VIII0度.此外,温度传感器的这个当前版本不防篡改和记忆机构发生故障时,应在传感器I.VIII0度,并再次冷冻旋转.我们的研究工作,目前正在研究开发I.种低成本的固态传感器,它是独立的传感器放置方向和不需要部署之前被初始化.
延伸的传感器来检测的几个阈值的违反,也许是通过使用具有不同熔点的复合电介质的另I.个重要步骤.这可能是有用的设计,检测不仅是上限临界温度,但其下限以及I.个传感器.例如大多数有机农产品将运输来优化温度0?I.0℃之间.食物会破坏如果保持很长I.段时间高于I.0℃,而且会破坏如果冻结低于0℃.开发具有优化的RFID的温度传感器范围内,扩展监视不同的温度阈值范围内,哪些是独立部署的包装材料因而将是I.个显著推动了建议的技术,因此是未来研究的重点.
II.外文原文
RFIDTagAntennaBasedTemperatureSensing
RahulBhattacharyya,ChristianFloerkemeierandSanjaySarma
AutoIDCenter
MassachusettsInstituteofTechnology
Cambridge,Massachusetts0III.IIIIX
Email:rahulb,?oerkemandsesarma@mit.edu
Abstract_Temperaturemonitoringisimportantinanumberof?elds,particularlycoldsupplychainapplications.Mostcommercialwirelesstemperaturesensorsconsistoftransceivers,memoryandbatteriestomaintainatemperaturetimehistorybutthisisexpensiveandallowsforlimitedsensordeployment.Inthispaper,weproposealowcosttemperaturesensorbasedontheparadigmofpassiveRFIDtagantennabasedsensing.AsimplemechanicalmethodtopermanentlyinducechangesinRFIDtagpowercharacteristicsuponexposuretotemperaturesgreaterthanathresholdispresented.Criticaltemperaturethresholdviolationscanthenbedetectedbymonitoringreceivedbackscattersignalstrengthatareader.Thefeasibilityoftheproposedhypothesisisexaminedviatheoreticalandexperimentalmeans.Itwillbeshownthatthissensingparadigmhasthepotentialtogreatlyincreasethepervasivenessoftemperaturesensingnodesandimprovesupplychainvisibilityandperformance.
I.INTRODUCTION
Temperatureisanimportantenvironmentalparameterinadiversevarietyof?eldssuchasinfrastructuremonitoring,chemistry,environmentalengineeringandcoldsupplychainmanagement.TemperaturemonitoringincoldsupplychainoperationsisparticularlyimportantasoutlinedbyO’Connor[I.]whopointsoutthattemperaturewithinatransportationpalletitselfvariesbyuptoIIIV%.FluctuationsinambienttemperaturehavebeenknowntoaffectthequalityofperishableproduceasoutlinedbyEstrada-Flores[II]andSchusteret.al[III].WithanestimatedIII0%ofperishablegoods[IV]spoilingintransitinsupplychainoperations,thereisadireneedforincreasedvisibilitywithoptimalgranularityusingsensingdevices.
Temperaturesensorsarerequiredtomonitorandrecordallcriticaltemperaturechangesintheenvironmentovertheperiodofdeployment.Thewidevarietyofwirelesstemperaturesensorsdeployedincommercialsupplychainapplicationsachievethisbyincludingappropriateonboardelectronicslikememoryandbatterywhichdrivesupthecostofthesensorunit.Asaconsequence,economicconsiderationsoftenprecludethedeploymentofthesesensorsonatrulypervasivescale.Ambienttemperaturelevelsareinferredbasedontheoutputofa?nitenumberofsensornodes,ratherthanbymonitoringeachlogisticunitpassingthroughthesupplychain.Itwouldthereforebeveryusefultohaveadedicatedsensormonitoringatthelogisticunitlevelsothatcriticaltemperaturestatechangescanbemonitoredandrecorded.Howeverinordertomeetthisobjective,thecostofthissensormustbesigni?cantlylow.Inthispaper,weproposethedesignofanultra-lowcosttemperaturesensorbasedonpassiveUHFRFIDprinciples,whichiscapableofloggingtemperaturechangesaboveacriticaltemperaturethresholdforuserspeci?edtoleranceintervals.
SectionIIdiscussestherequirementsoftemperaturesensorsandtheimpactoftheserequirementsonthescaleofdeploymentandmeasurement.InSectionIIIwediscussourparadigmofRFIDtagantennabasedtemperaturesensing,howitwouldmeettheserequirementsanddescribetheconstructionofaprototypesensorbuilttoexaminethishypothesis.SectionIVthenattemptstoexaminethefeasibilityofthisprinciplebydrawingcomparisonstothebehaviourofadipoleantennainamulti-layereddielectricmedium.SectionVoutlinestheresultsofexperimentationonthesensorprototypeusedtovalidatethishypothesis.Finally,inSectionVIwesummarizethecontributionandoutlinethescopeforfuturework.
II.CURRENTSUPPLYCHAINTEMPERATURESENSINGALTERNATIVES
Temperaturemonitoringisparticularlyimportantinthecoldsupplychainwheretemperatureshavetobearti?ciallymaintainedtooptimizetheshelflifeofthetransportedgoods.Forexample,orangejuicetypicallyproducedandpackagedinFloridaismobilizedviatransportationnetworkstodifferentcornersoftheUS.Inordertoprolongthelifeofthejuice,thetransportationmediumisrequiredtobeappropriatelyrefrigerated.Similarlyvaccinesandotherperishablemedicalitemsmayrequiretobeconstantlystoredatsub-zerotemperaturesinordertoremainpotent.
Itiscrucialtologthecriticaltemperaturestatechangesduringtemperaturemonitoringofanenvironment.Mostofthecommercialsensors[V][VI][VII]availabletodayarecustomizedforpreciselythisfunctionalityandarecomprisedofaseriesofdiscreteelectroniccomponentssuchasantennas,applicationprocessors,batterypowerunits,signalconditioningunitsandmemory.Unfortunately,thepresenceofdiscreteelectroniccomponentsdrivesupthecostofthetemperaturesensorandthismanifestsinareductioninthenumberoffeasiblesamplepoints,duetomonitoringbudgetconstraints.Lackofadequatemonitoringgranularity,couldinadequatelyrepresenttheambienttemperaturepro?leandthereforeitisimportanttomaintainanappropriatebalancebetweensensorfunctionalityandcost.
Withthisinmind,wehighlightthemostimportantfactorsrequiredoftemperaturesensorsinsupplychainmonitoring:LowCost:MostcommercialtemperaturesensingoptionsaretypicallyavailableinthepricerangeofO($II0)today.Thepriceofthesesensorspreventtheirpervasivedeploymentinsupplychainapplications.Forexample,itdoesnotmakeanycommercialsensetotaganorangejuicecartonpricedatabout$IIIwithadedicatedsensorwhichcostsO($II0).Thustemperaturemeasurementswiththissensorislimitedtoinferencemeasurementsfroma?nitenumberofnodesdeployedinthefreezerorrefrigeratorstorageunitwhichcontainsthousandsofsuchcartons.Thegranularityoftemperaturemeasurementsmaynotbesuf?cienttocapture?uctuationsintemperatureacrossthefreezerunit.Inordertoattaintrueubiquityinsupplychainoperations,thesensorsshouldbeultra-lowcostsoastoimprovethescaleoftheirdeployment.
AbilitytoMaintainState:Mosttemperaturesensorstodaylogatimehistoryofdatasothatthresholdtemperatureexposureinformationcanberecorded.Thequestionofwhetheracriticalexposurethresholdwasattainedismoreimportantthanthetimehistoryitself.Itisthusimportantthatthesensorsbeabletologexposureinformationatminimumcost.StandardizationofCommunicationProtocols:Therearenumerouspartnersinteractinginasupplychain.Inordertofacilitateseamlessdatatransferfromonepartnertoanother,thetemperaturesensor’scommunicationprotocolmustbestandardized.Itisimportanttopicksensorsthatconformtoaprotocolthathastheleastsetupcostintermsofhardwareandsoftwareequipment.
RFIDpresentsawellestablished,standardizedwirelesstechnologytoleverageforsensordevelopment.Somecommercial
alternativeshaveattemptedtoprovideRFIDbasedtemperaturesensingsolutions.Forexample,KSW[VIII]hasdevelopedanRFIDsmartlabelwithanintegratedtemperaturesensor.Similarly,Gentag[IX]hasdevelopedanRFIDtagICwithanintegratedtemperaturesensor.BoththeseapproachesleveragetheRFIDGenIICommunicationprotocol[I.0]andthusconformtoastandardizedcommunicationprotocol.TheKSWsensorhowevercostsaround$III-$I.0andcannotbedeployedfortaggingcheaperitems,suchasperishablefoodproducts,whichwouldaccountforasizablefractionoftheitemspassingthroughthesupplychain.SimilarlythedesigncostsassociatedwiththeGentagICwoulddriveupthecostofthecompositesensortagwhenevertheICisintegratedwithatagantenna.ThereisthusaneedtodrivethecostofatemperaturesensorevenlowerinordertoensureitsubiquitousdeploymentandinthispaperweexaminethedesignofatemperaturesensorusingtheRFIDtagantennabasedsensingparadigm.TherehasbeensomepriorresearchworkinusingtheRFIDtagantennaasasensor.Forinstance,Marrocco[I.I.]presentbasictheoryandsimulationsfortheperformanceofmultiportRFIDtagsonmaterialswithdifferentpermittivities.Similarly,Bhattacharyyaet.al[I.II]presentatagantennabaseddisplacementsensorbyrelatingdisplacementtoachangeinreaderthresholdtransmittedpowertopoweruptheRFIDtagICandthedifferentialbackscatterpowerreturnedfromanRFIDtag,whileSidenet.al[I.III]illustratehowordinaryRFIDtagscanbeusedashumiditysensors.Inrelatedwork,Marroccoet.al[I.IV]presentexperimentalresultsfortagsensordesignsalongthelinesofthetheoryandsimulationspresentedin[I.I.].Inthenextsection,wediscussthedesignofanRFIDtagantennabasedtemperaturesensorthathastheadvantageofbeingultralowcostandlonglastingwithanabilitytomaintaincriticalthresholdtemperatureexposureinformationusingnon-electricmemory.
Fig.I..TemperatureSensorInitialization
Fig.II.TemperatureSensorWorking
III.TEMPERATURESENSOROPERATINGPRINCIPLE
RFIDtagantennabasedsensing[I.II]isaparadigmfordevelopingultralowcost,longlasting,standardizedandaccuratesensorsleveragingtheUHFRFIDinfrastructure.
ThissensingapproachreliesonmappingachangeinsomephysicalparameterofinteresttoacalibratedchangeinRFIDtagantennacharacteristicsandthusRFIDtagpowercharacteristics.
TherehasbeenpriorresearchintothedevelopmentofRFIDTagAntennabasedsensors.Forinstance,Bhattacharyyaet.al[I.V]discussthedevelopmentofadisplacementsensorusingthefacttheRFIDtagperformancedegradesincloseproximitytometal,whileSidenet.al[I.III]demonstratehownormalRFIDtagscanbeusedtodetectrelativehumidityandmoistureingress,bymakinguseofthefactthatRFIDtagperformancedegradesinproximitytowater.
Inthisstudy,wedesignanRFIDtagantennabasedtemperaturesensorforsupplychainscenarioswherethetemperatureofthetransportedgoodscannotexceed0oCforperiodsoftimeexceedingachosentolerancelimit.Ifthetemperatureismaintainedabove0oCformorethanthetolerance,thesensoriscapableofloggingthisandconveyingthisinformationtothenextRFIDreaderunitthatinterrogatesthetemperaturesensor.TheoperatingprincipleofthesensorissummarizedinFig.I.andFig.II.
ThetemperaturesensorisdesignedasacompactcuboidaljacketmadeoutofplexiglasswithtwoRFIDtaginlayspastedontheoutersurfaceoftheplasticjacket,separatedbyaboutIII-Vcm.Theinsideoftheboxis?lledwithanaqueousmediumhavingadesiredmeltingpoint-forthepurposesofthisstudy,wechoosewaterhavingameltingpointof0oC.Analuminiumplate,capableofverticalmotionintheaqueousmediumislocatedbehindthetwotagssuchthatatanygiveninstantoneofthetwotagsisdetunedduetotheproximityofthemetalplate.Thesensorreliesonchangesinphaseoftheaqueousmediumandverticalmotionofthealuminiumplateundertheforceofgravitytorecordchangesincriticalstateasweshallillustrate.
A.SensorOperatingPrinciple
Wenowanalyzeindetailtheoperatingprincipleofthissensor:
Initialization:Thetemperaturesensorisdesignedtomonitorthetemperatureofgoodsinafreezerbelow0oC.Thesensoristhusinitializedbyfreezingtheicewiththemetalplatedirectlybehindthetoptag-whichweshallhenceforthrefertoasTagAasseeninFig.I..Thusintheinitializedstate,wewouldexpectbetterperformancefromTagB,bothintermsofdifferentialbackscatterpowerfromTagBaswellasthresholdtransmittedpowerrequiredtopowerupTagB,sinceTagAisdetunedduetothemetalplatedirectlybehindit.Ifduringsupplychaintransportation,thegoodsremainatorbelow0oCforalltimeslessthanthetolerancelimit,themetalplateremainsfrozenbehindTagA.Eachreaderunitinterrogatingthesensorwouldbeabletoverifythatthegoodsweremaintainedattemperaturesbelow0oCthroughouttransit.Fig.I.illustratestheattachmentofthesensortooneoftheitemsintherefrigeratorpriortotransportation.
EffectofHighAmbientTemperature:Ifthetemperaturesensorisplacedattemperaturehigherthanthetemperatureofthefreezerforperiodsgreaterthanthetoleranceinterval,theicewouldmeltandthemetalplatewoulddescendunderthein?uenceofgravitytoapositionbehindTagBasseeninFig.II.Thefactthatthetemperaturesensorhasundergoneastatechangecanthenbedetectedbythefollowingobservations:
–ReductioninReadRange:Thewatermoleculeishighlypolarandtendstoorientitselfsoastocanceloutanyincomingelectric?eld,aphenomenonthatisnotaseffectiveinthesolidicestate.ThusaboutVIII0%oftheincomingelectric?eldwouldbere?ectedattheplexiglass-waterinterface[I.VI].Thisismanifestedinareductioninsensorreadrangefortagpowermeasurementsfrombothtags.Thusifasensorarrivesatadistributioncenterwiththeaqueousmediumintheliquidstate,thereadrangeforbothtagsisdrasticallyreduced.
–DetuningofTagB:SincethemetalplatedescendsundergravitytodetuneTagBinsteadofTagA,achangeofstatecanalsobecharacterizedbyadegradationofperformanceofTagBrelativetoTagA-whichisindependentofwhethertheaqueousmediumisintheliquidorsolidstate.Incontextofthesupplychain,ifacrateoffrozenproduce,taggedwiththissensor,iskeptoutinthesunforacoupleofhours,theiceinthesensorwouldmeltandthemetalplatewoulddescendtodetuneTagB.Evenifthegoodsareplacedbackinthefreezer,themetalplatecontinuestodetuneTagB.Thusitispossibletopermanentlyrecordthefactthatthetemperaturesensorsufferedacriticalstatechange,irrespectiveofwhatstatetheaqueousmediumassumessometimeafter.
Fig.III.TemperatureSensorinElevation
TheplanandelevationofthesensorisdescribedinFig.IIIandFig.IVrespectively.
B.PrototypeDesign
Weexaminethefeasibilityofthissensingapproachviathedevelopmentofaprototypesensor(c.fFig.V).Theplexiglassjacketwhichhousesthemetalplateandthe
aqueousmediumisdesignedforthedualpurposeofI.)testingtheperformanceofdifferentkindsofRFIDtaginlaysandII)allowingfordifferentvolumesofwatersothattheice-watertransitiontimecanbecalibratedandcontrolled.
SelectionofJacketSurfaceArea:RFIDtaginlaysvarysigni?cantlyindimensions,basedonantennageometryfromaboutI.0cmXIIcminthecaseofanAlienSquiggle[I.VII]toaboutIXcmXIXcminthecaseoftheImpinjBanjo[I.VIII].Thusthesurfaceareaoftheplasticjacketwassetasl=I.VIcmandb=I.0cm(Fig.III)tobeabletoaccommodatemostkindsofinlays.ThecompletelistofdimensionsisoutlinedinTable.I.ReferencesshouldbemadetoFig.IIIandFig.IVtointerpretthedimensions.
EffectofVolumeofWateronToleranceInterval:
InFig.VIweplotobservedtimerequiredfortheicetocompletelymeltasafunctionoffourdifferentcontrolvolumesvaryingfromIV0mltoI.00mlofwaterasshowninTable.IIforaconstantoutsidetemperatureofI.VIIIoC.We?ndthatalinearbest-?tfunctionadequatelycapturestherelationship.
FromFig.VIweobservethatitispossibletodrawafairlyaccuratecorrespondencebetweentheexperimentallyobservedmelttimeandacontrolvolumeofwater.Forthepurposesofexperimentation,acontrolvolumeofVVml,correspondingtoatoleranceintervalofaboutI.VIIImins,ischosen.Ingeneralhowever,thevolumeofwatercanbevariedsoastomeetthespeci?edtoleranceintervalbyreferringtoFig.VI.
Fig.IV.TemperatureSensorinPlan
Fig.V.TheSensorPrototype
TABLEI
DIMENSIONSOFTHEPROTOTYPE
Dimension)Value(mm)
JacketLength(l)I.VI0
JacketBreadth(b)I.00
JacketDepth(d)VI
JacketMaterialThickness(s)I..III
TABLEII
VARIATIONOFICEMELTTIMEWITHVOLUMEOFWATER
Volume(ml)MeltTime(min)
IV0I.VI
VIVII0
VIII0IIVII
I.00III0
Fig.VI.PlotofMeltTimeasafunctionofdifferentcontrolvolumes
IV.THEORETICALANALYSISOFSENSORPRINCIPLE
Changesinbackgrounddielectricmaterialwillcertainlyin?uenceRFIDtagperformancebothintermsofthepowerbackscatteredbythetagaswellasthethresholdtransmittedpowerrequiredtopowerupthetag.ThesensorprincipleoutlinedinSectionIIIreliesonthedifferenceinbackscatterpowerlevelsduetothepresenceorabsenceofthemetalplatebeingsigni?cantenoughtobeabletodifferentiateastatechange.Inthissection,weproposeaquickmathematicalanalysistodemonstratewhetherthisisindeedexpectedtobethecase.
Thenetfar?eldduetoadipoleantennaontopofadielectricsubstrateandassumingTransverseElectric(TE)radiationisduetothesuperpositionoftheelectric?elddirectlyradiatedfromthedipole,andthecomponentoftheelectric?eldre?ectedbythedielectricsubstrate,.Thissuperpositionisgivenby:
(I.)
whereistheeffectivere?ectioncoef?cientduetoasystemofdielectriclayersandisthefarelectric?eldofthedipoleantennawithnobackgrounddielectricinfreespace.
Fig.VII.DielectricLayersinthemodel
Weconsiderthreecaseshere.Inthe?rstcase,thebackgrounddielectrictotheRFIDtagistheplexiglassjacketfollowedbythemetalplate.Sincethemetalplateisassumedtobeperfectlyre?ecting,wedonotconsideranysubstratebeyondtheplate.Inthesecondcase,thebackgroundmaterialisplexiglass,followedbyalayerofice,followedbyplexiglassasshownbytakingsectionsatA-AandB-BinFig.IIIandillustratedinFig.VII.Thethirdcase,isthesameasthesecondcaseexceptthattheicedielectricisreplacedbywater.
AsoutlinedinKong[I.IX],thecoef?cientcanbecalculatedforannlayersystemby
(II)
Wede?netheterminologyusedinEq.IIbelow:
Heremistheeffectivere?ectioncoef?cientofthesystemofdielectricscorrespondingtothethreecasesm=I.,II,IIIformetal,iceandwaterdielectricrespectively.
relatestothethicknessofthelayerinthedielectricsystem.
isthezcomponentofthewavenumberinlayern.
isthere?ectioncoef?cientbetweentheandlayersandisgivenby:
(III)
herewehave
(IV)
whererelatestothemagneticpermeabilityoflayeriandisthecomponentofthewavenumberinmediumj.Fornormalincidence,relatestotheelectricalpermittivityforthelayerj.
TABLEIII
RETURNEDPOWERLEVELSOFICEANDWATERRELATIVETOMETAL
DielectricPowerRelativetoMetal
IceIIVIII.VVIIdB
WaterIII..I.VIIIdB
SincethePoyntingpowerdensityis,wehave
(V)
whereAistheconstantofproportionality.AssumingEisthere?ectioncoef?cientsforcaseI.,IIandIIIwehavethedifferenceinpowerlevelsoftheiceandwaterdielectric,relativetothemetalplatedielectricas
(VI)
and
(VII)
PlugginginthevariousvaluesfordibyreferringtoFig.VIIandmakingappropriateassumptionsfortherelativedielectricconstantsforwater,iceandmetalweobtainresultsofpowerdifferencelevelsassummarizedinTableIII.Wenotefromthesecomputedvaluesthatthedifferenceinpowerlevelsforiceandwateraresuf?cientlylargerelativetometalasabackgrounddielectric.Furthermore,thereisarelativedifferenceofaboutVIIdBbetweeniceandwaterandthusallcriticalstateandphasechangesareclearlydiscernible.Thusthefeasibilityandvalidityofthissensorshouldcertainlybeinvestigatedviaexperimentationandthisisthefocusofthenextsection.
V.EXPERIMENTATIONANDRESULTS
Inordertoverifythesensinghypothesis,wesubjectthesensortotemperaturecyclesfromthefrozenstatetothefullymeltedstate.AcontrolvolumeofVVmlofwaterisselectedforexperimentation.ReferringtoFig.VI,weexpectthemelttimetobeaboutI.IXminutes.
TagreadrangeisakeyissuetobedeterminedinthisexperimentsinceRFIDtagperformancedwindlessigni?cantlyinthepresenceofwaterandmetal.Thesensor-RFIDreaderseparationisvariedfrom0.VmtoI..Vminstepsof0.Vm.Foreachseparation,thesensoriscycledfromthefrozenstatetothemeltedstateandthereceivedbackscatterpowerfromTagsAandB(cf.Fig.III)asafunctionoftimeisplotted.WeexpecttheicetocompletelymeltinaboutII0minutes,howeverthetestisconductedforaI.hourinterval.
RFIDReaderSystem[I.VIII]isusedintheexperimentation.ThereadertransmittedpowersettingissettothemaximumofIVWEIRP.ThetemperaturesensorintheinitializedfrozenstateisplacedbeforethereaderantennaatanenvironmenttemperatureofI.VIIIoC.ThereaderantennameasuresthedifferentialbackscatterpowerfromTagAand
BasafunctionoftimeforaI.hourinterval.ReadingsaretakenattherateofIIpersecond,averagedoverVI0secondsandreportedonaperminutebasis.Wenowdiscussthesensorresponseasafunctionoftimeforthethreedifferentsensorreaderseparations.
Fig.VIII.BackscatterPowervs.timeforsensor-readerseparationof0.Vm
A.Reader-SensorSeparationof0.Vm
Fig.VIIIplotsbackscatterpowerresponsefromthetwotagsasafunctionoftime.Initially,TagAhasthemetalpiecebehindit,whichdetunesthetagantenna,asaresultofwhichTagAgivesalowerbackscatterpowerresponserelativetoTagB.TagBinitiallyhasiceasthebackgrounddielectric,whichdetunestheantennalessseverely,asaresultofwhichTagBinitiallyhasamuchbetterbackscatterpowerresponserelativetoTagA(cf.Fig.I.).
Fig.IX.BackscatterPowervs.timeaccountingforcondensationmoisturebuildup
Astimepasses,theicestartstomeltandthebackgrounddielectricforbothtagsinvolvesincreasingamountofwater.Watermoleculestendtoorientthemselvessoastocanceloutincomingelectric?eldsandthuswaterdetunesthetagantennaandre?ectsmostoftheincomingelectric?eld,thusaffectingperformancemoreseverelythanice.Thusbetween0andI.0minutes,theperformanceofbothtagsstartsreducing.BetweenI.0andIIIVminutesweseethatthereisnoobservedresponsefromTagA.Weobserveduringexperimentationthatwatervapourhasatendencytocondenseonthesurfaceofthecoldplexiglassjacketandarbitrarilydetunethetags.ThecondensedwaternowpresentsoneadditionallayerinfrontofthetagwhereVIII0%oftheincoming?eldisscattered.Itispossiblethatthiscondensationtogetherwiththeproximityofthemetalplate,detunesTagAtotheextentthatthetagICchipfailstopowerup.
Inordertoverifythishypothesis,werepeattheexperimentremovingmoistureduetocondensationassoonasitbuildsup.TheresultsareseeninFig.IX.Fromthe?gure,weobservethatTagAalwaysgivesaclearresponseoverIIIVminutes,illustratingthatmoisturecondensationmustbeaccountedforwhiletakingmeasurementsfromthesensorsystem.
ReferringbacktoFig.VIII,itisobservedthattheicecompletelymeltsandtheplatedescendstodetuneTagBataboutIIIVminuteswhichisreasonablyclosetothepredictedI.IXminutes.ThusafterIIIVminutes,TagBgivesaworseperformancethanTagA.Wenoteaninterestingobservationhere.AfterIIIVminutesTagAisnolongerdetunedbythemetal,butitsperformanceisworsethanthatofTagBatt=0,sincethebackgrounddielectricforTagAiswater,whichdetunesthetagantennamoreseverely,ratherthanice.Toverifythis,werefreezethesensorwiththemetalplatebehindTagB.AsseenfromFig.IX,theresponseofTagAthenimprovestolevelscomparablewithTagBatt=0.
Ifweattempttocorroboratetheexperimentalresultswithourtheoreticalanalysis,weobservethatthetrendspredictedbythetheoreticalanalysisarecon?rmedviaexperimentation.ThepowerlevelswithdielectricbackdropoficeisonanaverageaboutVIIdBmbetterthanwithabackdropofwaterasseenbycomparingthepowerlevelsofTagBatt=0withthoseofTagAbeyondt=III0inFig.VIII.Similarly,thereisasigni?cantdifferenceofaboutI.0dBmbetweenabackdropofmetalvisa-visice.AlthoughnotashighasthetheoreticallycomputedIIVIIIdBm,theassumptionoftherebeingasigni?cantdifferenceinpowerlevelbetweenthestatesiscertainlyveri?edinpractice.
Inordertovalidatetherepeatabilityoftheobservations,thetemperaturecycletestwasrepeatedthreetimesandineachrun,repeatableresultswereobtained.
B.Reader-SensorSeparationofI.andI..Vm
Fig.VIIIshowstheresponseofthesensorasafunctionoftimewhentheseparationwasincreasedtoI.m.Thetrendsofthecurvesremainthesamealthoughtheabsolutevaluesofthepowerlevelsdecreaseslightlyduetotheincreaseddistancebetweenreaderandtag.ItisinterestingtonotethatatI.mdistances,oncethemetalplatedescends,TagBdoesnotgetpickedupbythereader.Thepowerdensityreachingthetagovertheincreaseddistanceaftermismatchlossesduetothewaterandthemetalisinsuf?cienttopowerupTagB’sIC.Similartrendswereobservedwhenthereader-tagseparationwasincreasedtoI..VmasshowninFig.I.I..Asweobservefromthe?gure,thereareseveralmissedobservations,indicatingthatthebackscatterpowerreadingsfromthesensoraredif?culttodetectatthatrange.
Fig.I.0.BackscatterPowervs.timeforsensor-readerseparationofI.m
Fig.I.I..BackscatterPowervs.timeforsensor-readerseparationofI..Vm
Asinthepreviouscase,thesetestswererunthreetimesaswellandconsistentresultswereobtained.Fig.I.IIforinstancedemonstratesthreerunswhenthesensor-readerseparationisI.m.
Fig.I.II.ThreeRunsforReader-SensorSeparationofI.m
Fig.I.III.Effectofdifferentmaterialofdeployment
C.EffectofMaterialofDeployment
Itisalsoimportanttoconsidertheelectromagneticin?uenceofthematerialonwhichthesensorisdeployed.Theexperimentswereconductedbyattachingthesensortoahollowcardboardbox,howeveritisnecessarytoconsidertheeffectofothermaterials-particularlythesocalledRFunfriendlymaterials.Alotofpackingcratesandcontainersaremetallicandinordertosimulatethisthesensorisattachedtoametallicplateandtheexperimentisrepeatedfora0.Vmsensor-readerseparation.TheresultsareshowninFig.I.III.Aswecanseethepresenceofmetalsigni?cantlyalterstherelativepowerdifferencesbetweenthetwotags.Infact,themetalseemstoboostthestrengthofthebackscatterpowersignal.WenotethatthemetalplateislocatedatadistanceofaboutIIcmfromthetags.Itispossiblethatthebackscatterpowerfromthetagandthere?ectedsignalfromtheplateareinterferingconstructivelyatthereaderandthismayaccountfortheboost
insignal.Thusmaterialofdeploymentisanimportantfactortobeconsideredandwillbeaddressedinfutureworktoenhancethesensorperformance.
VI.CONCLUSIONS
Inthispaper,weproposeadesignmethodologyandverifytheworkingfeasibilityofanultralowcostwirelesstemperaturesensorwhichleveragestheRFIDinfrastructure.Whilemanyofthesensorscommerciallyavailableinthecoldsupplychain,boasthighprecisionandaccuracyofmeasurementwithfeaturessuchasmemorylogging,theirrelativelyhighcostprecludestheirpervasivedeploymentincoldsupplychainapplications.
Ourapproachemphasizestheultra-lowcostofmanufacturebutrecognizestheassociatedtrade-offsthiswouldentail.Tagantennabasedtemperaturesensingwillnotproviderealtimeupdatesorlogatimehistoryoftemperaturemeasurements.Howeverthissensorwillprovidetheuserwithbinarystateinformationindicatingwhethercriticaltemperaturethresholdswerereachedduringtransportationallowingtheusertomakeadecisionofwhethertokeepordiscardthegoods.Sincethestateofmostgoodsareexaminedatthedistributioncentersinasupplychain,ratherthanintransit,wearguethatthiskindofsensingmightinfactbeadequateformonitoringmostitemssuchasorangejuicecartons,milkandmeatproductsinsupplychainapplications,sincethestateoftheproductswillbeautomaticallymonitoredassoonastheproductspasspastanRFIDreaderatthedockdoorofadistributioncenter.Furthermore,sincemanysupplychains
haveanRFIDinfrastructurealreadyinplace,thissensingparadigmhastheadvantageofleveraginganalreadywellestablishedcommunicationinfrastructure.Thuswithitsultra
lowcostandabilitytoconveycriticalstatechangeinformationuponinterrogation,weexpectRFIDtagantennabasedsensingtobeveryrelevanttoalargemarketshareofcoldsupplychainapplications.
Whilethedesignmethodologywasdemonstratedinthispaper,thereissigni?cantresearchthatrequirestobedoneinthisareatoextendthetechnology.WenotethatthelowerTagBisconsistentlynotreadforreader-tagseparationsofmorethan0.Vm.Itisdif?culttodifferentiatethisfromafailureofTagB.ArelativeresponsefrombothtagsAandBwouldbeamuchbetterindicatorofstatechange.Itisthusnecessarytodesignatagwithcustomantennageometrythatisoptimizedforperformancenearwaterandmetalsothatthereadrangeofthesensorisboostedtomorethan0.Vm.
ThecurrentprototypeimplicitlyassumesthatthesensorwouldbedeployedonpackagingmaterialthatisRFfriendly.Inpractice,manytransportationpalletsaremetallicandtheperformanceofthesensorisaffectedbythis.Modifyingthesensordesign,perhapsbyincludingametalbackplane,wouldmaketheperformanceindependentofthematerialofdeployment.
Condensationofmoistureonthesurfaceofthesensorisanotherissuethatmustbetackled.Evenforreader-sensorseparationsofabout0.Vm,thebuildupofmoisturecandetunethetagandre?ectincidentelectric?eldstotheextentthatthetagisnolongerpickedup.Eliminationofmoisturebuildupbyuseofappropriatehydrophobicsurfacesthatshedwateroncontactisanotheravenuethatmustbeexploredinthesensordesign.
Therequirementthatthissensorbeinitializedbyfreezingthemetalplateinpositionbehindthetoptagisonepotentialimpedimenttothecommercialdeploymentofthissensor.Althoughitisnotdif?cultforsupplychainpartnerstoinitializebatchesofthesesensorsinindustrialfreezers,itisoneadditionalstepthatdetractsfromthelow-cost,pervasiveappealofthissensor.Furthermore,sincethememoryeffectistriggeredbygravity,itisnecessarytoplacethesensoronthesideoftheshippingcontainer.ThusanunderlyingassumptionisthatthegoodsarenotrotatedorturnedbyI.VIII0degreesduringtransit.Furthermore,thiscurrentversionofthetemperaturesensorisnottamperproofandthememorymechanismbreaksdownshouldthesensorberotatedbyI.VIII0degreesandrefrozen.Ourresearcheffortsarecurrentlylookingintodevelopingalow-costsolidstatesensorthatisindependentofsensorplacementorientationandwhichdoesnotrequiretobeinitializedbeforedeployment.
Extendingthesensortodetecttheviolationofseveralthresholdsisanotherimportantstepperhapsviatheuseofcompositedielectricshavingdifferentmeltingpoints.Itmightbeusefultodesignasensorthatdetectsnotonlyaupperboundcritical
temperaturebutalowerboundaswell.Forexamplemostorganicproducewouldbeoptimizedfortransportfortemperaturesbetween0andI.0C.ThefoodwouldspoilifmaintainedforalongtimeaboveI.0Cbutwouldalsospoiliffrozenbelow0C.DevelopinganRFIDbasedtemperaturesensorhavinganoptimizedrange,withextensionsformonitoringdifferenttemperaturerangethresholdsandwhichisindependentofthepackagingmaterialofdeploymentwouldthusbeasigni?cantboosttotheproposedtechnologyandisthusthefocusoffutureresearch.
REFERENCES
[I.]M.CO’Connor,RFIDJournalArticle:Coldchainprojectrevealstemperatureinconsistencies,"
http://www.r?djournal.com/article/view/IIVIIIVI0/I.,DecemberI.,II00VI.
[II]S.Estrada-Flores,Acousticmeasurementofavocado?rmnessduring
lowtemperaturestorage:Effectofstoragetemperature,"International
ConferenceonQualityinChains.AnIntegratedViewonFruitand
VegetableQuality,pp.IIIIIIII–IIIIIIX,II00III.
[III]E.WSchuster,S.JAllenandD.LBrock,GlobalRFID:TheValueofthe
EPCGlobalNetworkforSupplyChainManagement,"SpringerVerlag,
II00VII.
[IV]A.DadaandF.Thiesse,SensorApplicationsintheSupplyChain:The
ExampleofQuality-BasedIssuingofPerishables,"LectureNotesin
ComputerScience,vol.IVIXVII/II00VIII,pp.I.IV0–I.VIV,II00VIII.
[V]EvidenciaTemperatureSensingSolutions,ThermAssureRFTemperatureSensor,"http://www.evidencia.biz/products/acti-tag.htm.
[VI]PultronicsWirelessDesignSolutions,TechnicalLiteratureonWarehouseTemperatureMonitoringSolution,"
http://www.pultronics.com/warehousetemperature.htm.
[VII]C.S.T.Solutions,Coldtrakcoldmonitoringsolutions,"
http://www.supatrak.com/cold-chain-monitoring.html.
[VIII]RFIDNewsArchives,NewLost-CostTemperatureSensor,"
http://www.r?djournal.com/article/view/IIVIII.
[IX]UsingRFIDNewsArchives,Low-CostTemperatureSensortoaddtoRFIDInlays,"
http://www.usingr?d.com/news/read.asp?lc=wIIIIIVVrxVIIVVzo.
[I.0]TheEPCGlobalStandards,ClassI.GenerationIIUHFAirInterface
ProtocolStandard,"http://www.epcglobalinc.org/standards/uhfcI.gII/,
II00VIII.
[I.I.]G.Marrocco,L.MattioniandC.Calabrese,MultiportSensorRFIDs
forWirelessPassiveSensingofObjectsBasicTheoryandEarlyResults,"
IEEETransactionsonAntennasandPropagation,vol.VVINo.VIII,pp.
IIVIIXI.–IIVII0II,II00VIII.
[I.II]C.Floerkemeier,R.BhattacharyyaandS.Sarma,BeyondtheIDin
RFID,"II0thTyrrhenianInternationalWorkshoponDigitalCommunications-aCNITConference,II00IX.
[I.III]Johan,S.andXuezhiZengandUnander,T.andKoptyug,A.and
Nilsson,H.-E.,RemoteMoistureSensingutilizingOrdinaryRFID
Tags,"Sensors,II00VIIIEEE,pp.III0VIII–IIII.I.,Oct.II00VII.
[I.IV]G.MarroccoandF.Amato,Self-sensingpassiver?d:Fromtheoryto
tagdesignandexperimentation,"MicrowaveConference,II00IX.EuMC
II00IX.European,pp.00I.–00IV,IIIXII00IX-oct.I.II00IX.
[I.V]R.Bhattacharyya,C.FloerkemeierandS.Sarma,TowardsTagAntenna
BasedSensing-AnRFIDDisplacementSensor,"IEEERFIDII00IX
ConferenceProceedings,pp.IXV–I.0II,II00IX.
[I.VI]D.M.Dobkin,TheRFinRFID:PassiveUHFRFIDInPractice,"
NewnesPublishing,II00VII.
[I.VII]AlienTechnologies,AlienTechnologyProducts,"
http://www.alientechnology.com/products/index.php,II00IX.[Online].
Available:http://www.alientechnology.com/
[I.VIII]ImpinjRFIDProducts,ImpinjUHFGenIIRFIDProducts,"http://www.impinj.com/products/default.aspx.[Online].Available:
www.impinj.com
[I.IX]J.AKong,ElectromagneticWaveTheory,EMW,Ed.JohnWileyand
Sons,I.IXVIIIVII.
版权保护: 本文由 hbsrm.com编辑，转载请保留链接: www.hbsrm.com/lwqt/wxzs/218.html
