Very Low-Cost Sensing and Communication.pdf
采用双向发光二极管实现廉价传感和通信外文文献翻译、中英文翻译
收藏
资源目录
压缩包内文档预览：
编号：91869287
类型：共享资源
大小：1.54MB
格式：ZIP
上传时间：2020-08-10
上传人：QQ14****9609
认证信息
个人认证
郭**（实名认证）
陕西
IP属地：陕西
12
积分
- 关 键 词：
-
采用
双向
发光二极管
实现
廉价
传感
通信
外文
文献
翻译
中英文
- 资源描述：
-
采用双向发光二极管实现廉价传感和通信外文文献翻译、中英文翻译,采用,双向,发光二极管,实现,廉价,传感,通信,外文,文献,翻译,中英文
- 内容简介：
-
第12页 共12页设计（论文）报告用纸采用双向发光二极管实现廉价传感和通信作者：Paul Dietz,William Yerazunis,Darren Leigh摘要：一个创新的单片机接口电路可以通过只用一个LED来交替发射和接收光信号，两个数字I/O引脚和一个简单的限流电阻来实现。这种技术首次被应用于创造一个智能照明系统，该系统只用简单的LED来做光源和光传感。然后，我们做出几个只用LED作为通用无线数据串端口的设备。该原理的一个重要应用是每个连接到单片机的LED都可以被认为是一个无线双向通信端口。我们把这种技术看做是对“最后厘米问题”的解决方法，因为它允许不同设备之间通过最小的设计改动来简单，廉价地通信。1. 说明LED，也称发光二极管，是最通用接口部件中的一种。它们不同的应用包括数码显示、闪光灯、液晶背光灯、煞车指示灯、交通信号灯和普遍的电源指示灯。由于LED被普遍用于光发射器，所以很容易忘记它们从根本上是光电二极管和光探测器这一点。虽然LED不能最好地用于光探测，它们仍是非常有效。在固态光发射和探测之间的可交替性在20世纪70年代被Forrest W. Mims广泛传播，但还是很大程度上被LED用户所遗忘。1.1 用LED探测环境光 最近，我们已经调查广泛用于消费者音像设备的远程红外遥控的提高方法。立即感兴趣的一块是用于许多远程遥控的按钮照明。为了打开背光，你必须按下按钮，在黑暗中，几乎是不可能找到该按钮。我们下决心要改善这种情况。 我们的第一个解决方法是用一个电容接近式传感器（类似于在【4】中描述的）来打开远程遥控背光。不幸地，每次打开背光实质上都是在消耗电池寿命，不仅仅是因为用户经常连续地远程遥控，而且有时是在良好光照下操作的，这时背光是不需要的。虽然加上了个模式开关，这并不比起先的情况要好。 明显的步骤是增加一个光检测器，只当有需要时才 打开背光。硫化镉光电池不贵，但给电池增加一光学路径会增加成本和机械设计的难度。回想起LED的感光特性，我们决定研究用背景灯LED本身作为光探测器。为了这个目的，我们设计了一个简单的电路，相对于传统的LED驱动电路，该电路只需要一个额外的单片机I/O引脚，而不需其他部件。 简单LED发射器/接收器的成功鼓舞我们考虑其他应用。例如，通过正向偏置（发射）和反向偏置(传感)之间的快速切换，可以创造出一个基于LED的常亮电源，但在实际上定期测量环境光照，利用这来自动调节LED的亮暗。图6所示的是我们的示例设备，用一个电容传感器来检测设备是否有操作，LED传感器/发射器来提供背光。1.2 LED通讯系统：双向LED通讯 环境光测量有许多应用，该技术的一项有趣应用是LED之间面对面地来回传输数据。我们称其为“LED通信”。我们已经研制出一个模型，该模型可以让LED在几厘米的距离内双向数据通信。一个LED通讯系统的可能应用是代替RFID系统（如【5】）进行支付授权和存取控制。为了测试可行性，我们已经创造一个廉价、钥匙扣大小的设备，该设备被称为信息采集器，可以接收、存储个传输数据。不像RFID系统，信息采集器支持真正的对等通信，允许新功能，例如，设备之间数据的直接转移，而不需要特殊的读取装置。 这种基于LED数据通信的想法是有意义的。每个连接到单片机的LED可以认为是一个无线通信端口。相比其他短范围的无线技术，例如，红外【7】和蓝牙【8】，LED通信范围更短，传输数据的速率更慢。但LED通信可以节约成本，在很多情况下，甚至免费。这是因为LED通讯系统本质上是用已存在的硬件通过小改动而形成的一种软件接口技术。在用传统技术成本较高的情况下，LED通讯系统可以代替其实现通信功能。在许多消费应用，电源灯现在可以成为一维修端口，用来读取服务信息或上传新固件。把两个手机的屏幕挨近，可以互相传输联系方式。对于汽车，标准，昂贵的服务连接器可以省略，所有数据可以通过“引擎检测”灯来传输。（汽车用户甚至可以用信息采集器来获取汽车故障记录，然后在服务预约前把它传输到服务中心，确保当汽车送入时，合适的工具和备用部件能立即应用。）这里有许多可能的应用。在下面正文中，我们描述了基本的LED双向单片机接口电路和它在智能背光的应用。然后我们对LED通讯系统、信息采集器和各种各样的应用进行充分描述。2. 双向LED接口小电流正向通过时，LED以相当窄的频率带发光。因为伏安特性曲线呈指数形式，很难足够准确通过控制LED上的电压来获取所需的电流。必须想办法限流。如图1所示，在分离系统中，通常加上一个电阻。由于大多数单片机I/O引脚可以接收电流，在图1所示的配置是通用的单片机驱动LED。 图1 典型LED驱动示意图 LED是一种对在发光上下波段的光很敏感的光电二极管（不包括任何有色塑料包装的影响）。在反向偏压条件下，LED相当于一个电容和一电流源并联(见图2，【3】)。这种光电流就是我们想要测量的。一种基于LED的光电探测器的廉价做法是把阳极接地，阴极连接一个高电平驱动的CMOS I/O 引脚。图2 LED反向偏压用于光电探测 二极管反向偏压，电容充电。接着，转换I/O引脚到输入模式，这时允许电容器放电低于数字输入阀值。通过测量花费的时间，我们得到光电流的测量方法和入射光的总数，如图3所示。图3 LED用作光电探测器 如图4所示，图1和图2的电路可以联合起来，创造出普通的LED双向单片机接口电路。这和图1的电路是一样的，除了电阻被两个I/O引脚间的LED代替。图5显示的是引脚如何驱动两个方式。图5a显示的是“发光”模式，这时电流正向流过，LED亮；图5b显示的是“反向偏压”模式，这时电容充电，准备测量系统。如图5c所示，实际的测量在“放电”模式中进行。由于流过CMOS引脚的电流非常小，低阻值的限流电阻对电压的影响非常小。像之前那样，我们只是简单测量电容放电低于阀值所花费的时间。结论是可以用一个简单的电路来转换发射和接收光的模式。 由于电路，改动需要提供双向通信功能，只包括一个额外的I/O引脚和印刷电路板（在设计期内提供，不需要额外的硬件成本）我们称增加该硬件功能到设备本质上是免费的，当然，为了完成这项工作，软件和CPU是必须的。图4 双向LED接口示意图 a）发光 b) 反向偏压 c) 放电图5 LED发射和检测光 把它和增加红外【7】（大约$7）和蓝牙【8】（超过$10）的成本相比，由于转换电平和静电放电保护电路的要求，使用如此简单的机械式连接器会花费几美元。已存在的LED用于通信也能节约制造成本，因为昂贵的塑料模型不需要改变以适应专门的红外收发器天线或物理连接。3. 智能背光智能背光是双向LED电路的一个应用。如前所述，智能远程控制背光的想法是在用户按下按钮之前打开背光。为了保持电力，我们希望只在足够暗的时候才打开背光。图6 自动背光原型为了验证这项功能，我们造了如图6所示的原型，完整图解如图7所示。该电路用一个电容接近式探测器来判断操作状态。虽然基本的电容测量电路和在手机缓冲区【4】用的一样，我们通过处理数据来寻找积极操作（电容变化），而不是简单的处理（电容增加）。许多用户即使在不积极用遥控时仍会持续使用，所以积极操作的探测对延长电池寿命有着决定性的作用。当然，一旦用户再次希望积极使用遥控，任何操作都可以打开背光。图7 自动背光示意图 智能背光功能如下：单片机周期性工作，测量电容。如果没有探测到积极操作，单片机不工作。否则，用LED进行光测量。如果房间很暗，打开背光至少两秒。每次操作被探测到，背光计时器重新设定，背光两秒。 由于远程遥控已经包含低端单片机，增加这项功能不会花费很多成本。接近电极可以是印刷电路板的一部分，除去特殊用途的需要。如果有多余的I/O引脚可用，唯一额外的部件是一个简单，廉价的电容，用于电容传感器。 也许你会问是否持续运行探测器会影响到电池寿命。事实上，电路会消耗电源的电力。该模型用一个简单，CR2032型号的，硬币大小的手表电池就可以连续运行6个月。远程遥控通常用AAA或AA的电池，该电池比硬币大小的电池有更强的蓄电能力。所以相比电池的自我放电，电源的损耗是无关紧要的。4. 双向通信协议在我们首次的智能背光的实验中，我们经常用基于LED的背光灯来测试光探测电路。这告诉我们LED之间的通信是可行的。我们想出一个简单的测试方法，用两个如图8所示的，相同的通用带有RS-232接口的PIC单片机。图8 LED双向通信 这些测试板利用一种简单的数据传输协议，其允许两个异步设备相位互相自锁，双向交换解调的脉冲带。对该协议的基本解释是两个设备的LED互相闪烁。短闪烁意味着“0”或space状态，长闪烁则是“1”或MARK状态。协议开启时，设备进入空转循环，发射一段1毫秒的光脉冲，紧接着是一段4毫秒的接收期。在接收期间，设备进行了40次光测量，每次耗时100微秒。这些光测量只提供一小部分解决方法，即是否入射光在数字I/O接口阀值（正常1.5V）上下波动。在100微秒接收期内，只有当正常室内光照在LED，电容放电，光电流低于阀值。如图9所示，示波器轨迹显示了在几个正常光照条件下进行光测量所得LED阴极的电压变化。竖直刻度是1V/格，水平刻度是100微秒/格。电容最初电压是5V，然后放电。注意电压始终不低于阀值，所以单片机引脚总处于“1”。图9 正常室内光照下的几组光测量 图10是同样步骤下的示波器轨迹，但是由另一个LED照明的。在测量期内，电容完全放电，I/O引脚电压低于阀值，处于“0”。在成功的“光照”下，空转循环至少持续两个测量周期。在这一点上，设备确保从另一相似设备来的入射光脉冲能被探测到，1毫秒的空转循环开启到4毫秒的接收期结束，转到一个更加快速的同步回路，描述如下。图10正常LED光照下的几组光测量 在同步回路期间，入射光仍是保持1毫秒，但紧接的是一系列的100微秒的光测量。在同步回路内，单片机会在另40个测量完成后结束测量，或当光脉冲的边缘被探测到时。当在有光照和10次没有光照的来回背靠背测量中，边缘才有可能被发现。同步回路内部的执行模式包括一个设备的LED亮1毫秒，然后所有LED熄灭1毫秒，紧接着另一设备的LED亮1毫秒，最后所有的LED熄灭1毫秒。即使所有的设备的时钟频率误差高达25%，它们仍能同步。名义上的同步回路脉冲率是250HZ，占空度为25%。图11显示的是两个设备在同步回路的示波器轨迹，互为光启动脉冲。注意它们的时钟是完全独立的，所有的同步都是通过LED和基本协议完成的。 在通信期间，数据位以异步形式传输。1毫秒的光脉冲表示MARK，0.5毫秒光脉冲表示SPACE。伴随MARK位传输，系统正常空转（数据传送回路和同步回路的软件是一样的）。在数据传输期间，以SPACE位作为起始位，紧接的是8位的数据位，最后是MARK位作为结束位。这和通用8-N-1 RS-232是一样的。图11的上面轨迹显示的是在设备空转状态下的数据脉冲轨迹，发送所有的MARK脉冲。下面轨迹显示的是设备发送窄SPACE和宽MARK脉冲。图11 同步操作下设备间的示波器轨迹 为了给光脉冲译码，接收设备保持对每一次“光照测量”下同步回路的执行次数。如果“光照测量”的次数是7次或低于7次，记为SPACE.如果是8次或高于8次，记为MARK。执行通常异步译码（去掉SPACE起始位和MARK结束位）。8位数据才对应用程序有用。一个简单，更高水平的电路允许误差存在和订正。LED通信测试步骤进行非常顺利。下层协议可以在各个方向以大约250位/秒的速率传输数据。单片机缓冲数据，然后以38400 bps的速率连接到主机。在大约3厘米的范围内的数据转换是强的。因为我们所用的LED有一个相当窄的光束角，它们只允许大约20度的误差。不像许多其他协议，该系统可以高度抵抗时钟速率误差。并不是所有的设备都有准确的晶振。即使时钟速率误差上升到25%,通信仍不受干扰。相反的，RS-232数据连接的误差超过5%经常会引发问题。在本应用中，廉价，不稳定的PIC设备内部振荡器实际上是有利的。即使两个设备同时调到相同相位，它们也会很快脱离同步，这时边缘将会探测到。设备会快速同步，交替闪烁。我们的设备在50微秒内会发生这种情况。如果两个LED通信设备都有高度稳定的定时器（或两个设备从同一来源得到时钟），将有必要插入一个振动源（也许是基于设备序列号的无用信息）到空转循环，确保两个脉冲信号能完全脱离，边缘能探测到。这个基本协议超过平衡脉冲协议的一项额外特征，例如曼切斯特代码，是LED可以为空转状态，同步状态和数据传输状态提供可见的提示。当准备传输数据时，光变亮（基于快速的脉冲重复率），在数据传输期间，光变暗（基于短0.5毫秒SPACE脉冲对没数据的1毫秒MARK脉冲）。5. 信息采集器s为了作为可行的信息便携储存或自动化，我们设计出一个被我们成为信息采集器（信息管道）的设备。像眼药水滴管那样，信息采集器可以吸收一小部分信息，分析信息，然后放出所需的信息。不像眼药水滴管那样，信息采集器可以不破坏性地，反复放出同样的信息。信息采集器可以用在用户希望由于经济或实际原因而没有可行的用户接口的设备之间传输数据。这也许是因为数据转移有时发生的太碰巧而不能添加一显示器和键盘到设备，即首次诊断安装信息。信息采集器可在设备和另一有用户接口的设备间传输数据。信息采集器的可行信息应用方面和mediaBlocks【9】的影响是相似的。主要的区别是mediaBlocks不能分析信息；它们只作为穿梭于网络中的信息代号。信息采集器本身可以分析信息和用作网络的一部分。信息采集器硬件包括一个小型印刷电路板，一个简单的按钮开关（唯一的用户输入），一个PIC16LF628单片机，一个LED（作用：数据输入，数据输出和用户输出），一个3V锂电池，一个电容和两个电阻。这里有5个额外的焊接带，以便为了实验可以增加新的部件。整个装备比大多数汽车警报远程遥控要小和便宜，且包含的部件少。一块产品应该花费不到1美元，超过类似的LED背光灯。信息采集器模型也配备了一个电路程序连接器，可以让我们把代码下载到单片机和改变设备的性能。我们也设计出一小块适配器板，可以把连接器装到单片机标准RJ-11电路调整模块上。图12所示的是两个信息采集器。下面那个有适配器，底部的可见的大塑料块是电池座设备最大的部件。在信息采集器的后尾，我们留下了一块1厘米的空白以便钻洞，然后和匙链相连。图12 一对信息采集器，其中有个有适配器默认的信息采集器性能规划是一个信息滴管。为了把信息吸入信息采集器，用户按下按钮两次，然后紧按住；然后，信息采集器会吸入任何给它的信息流并内部储存。当代码载入后和记录完成时，新颖的闪烁形式开启。过早的释放按钮会阻碍进程。为了防止信息流出信息采集器，用户按下并保持按钮；反复发射数据，大约1次1秒。这模式就像一个简单的闪光灯。这是这样设计的：信息采集器完成用作一个小型的闪光灯。设备里的锂电池可以允许10小时连续使用。当信息采集器不传输或接收数据时，单片机进入休眠模式。这降低了整个系统的功率要求，防止电池的泄露电流，延长了电池寿命。6. 作为智能键的信息采集器我们对信息采集器的目标之一是用作智能、可编程的键。虽然许多其他技术可用在智能键（RF、RFID和card-keys等等），LED通信有其独特的优势。第一，它不需要任何物理连接，所以不像一些card-key系统有机械磨损。第二，不像RF系统，它具有方向和短范围性，所以用户可以完全控制什么该被解锁。这允许一个简单的键用于不同的锁，而没有解锁错误的可能，因为它是相近的。第三，LED通信从根本上是双向作用的，允许挑战/反应和加密协议的应用，这使得键很难被复制或盗窃。第四，LED的可见特性允许一些用户接口。至少，用户可以简单地区分设备是否有操作和电池是否没电。第五，比起keycard和RFID读取，LED通信读取更简单，更廉价。这在锁的数目和健的数目是同样的序列的情况下是很重要的。第六，也许是最有趣的优势是LED通信能够点对点通信。任何LED通信设备可以传输信息或授权给另一个LED通信设备（如果应用软件允许）。在这种情况下，拥有标准“吸/挤”性能程序的信息采集器可以载入解锁代码，并传输给更多的信息采集器。这种传输授权的能力是完全唯一的，并不是智能卡或RFID所拥有的。为了演示信息采集器作为智能键的这项应用，我们添加了一个弹簧继电器和一个外部电源，把它嵌入安全系统，在这里锁码和解码。信息采集器的LED指向大厅的玻璃窗。信息采集器性能程序不工作直到它收到适当（秘密）的要求，当它接到该要求，激活继电器开门5秒。图13显示的是测试步骤。然后，我们给一个信息采集器载入正确的代码，并（像预期的那样）把它用于“开门”。充分应用LED通信对等的性能优势，我们把解锁代码传输给其他同样用于“开门”的信息采集器。图13 信息采集器用作关门（左）开门（右）7. 信息采集器、认证和安全在某些应用方面，传输信息或授权的对等能力是必要的。在其他应用，例如财务和其他安全处理，认证和转让一样重要，并授权的不可控制转移必须防止。一个信息采集器可编程特性的副作用是不能保证另一个设备会遵守“不可向前”的数据标签，该标签也许会装入到某种应用中。认证的不可转移性和身份认证的不可伪造性都是些伴随着微妙的困难程序。一个对如此高约束情况的解决办法将包括硬件、软件和超出范围的加密技术。 然而，简单的加密是可能的，并用于信息采集器处理保护。使用的单片机有充足的电力来完成通用对称密码算法。这些要求发送器和接收器拥有同样的密匙，所以任何两个设备之间的通信必须提前成型。信息采集器有足够的记忆来分析许多对称密码键，并能和许多其他设备进行对话。 零知识证明和公共键（或不对称）密码学能够让信息采集器安全地身份认证和任何设备通信。没有任何秘密是必须的。不幸的是，所有为这些要求广泛计算和数据通信（虽然仅在现代快速的工作环境中有效）的算法在信息采集器非常有限的算法环境中是不可能的。小型锂电池只有足够的电力让4MHZ的单片机运行几小时，所以一个需要在几秒钟内完成的算法会耗尽信息采集器整个电池寿命。我们现在正在研究零知识认证算法，该算法只需要有限的可用进程。有限数据率和信息采集器电池寿命的副作用是在电池寿命中，(大约106字节)全部的数据将被保留在廉价永久性存储器中。这允许一次性密码加密的使用，要么为了安全，要么为了有限单片机的使用。8. 每个LED都是通信端口虽然今天几乎每个电子设备都包含单片机和（理论上）有足够的能力和类似的设备通信，通信信路的花费经常阻碍两个靠近的设备进行通信。这就是“最后厘米问题”。有了LED通信技术，每个LED都可以变成一个潜在的通信端口。这拓宽了应用，因为在基于单片机的设备中，LED广泛用于电源指示。指示器通常不是直接连到电源，而是和单片机相连，如此小型用户接口是可行的。伴随适当的修改，指示器可以用于和信息采集器或其他LED通信设备通信。这种在有用户接口的设备和没有的设备之间，廉价、简单的数据传输的能力将允许设计师添加更多的性能到廉价产品中。又小又便携的产品可以装入用户交互接口，大型的设备也许就会要求用户用类似信息采集器的设备进行数据来回传输。以下是几种我们考虑的LED通信应用：1. 现代CRT监视器的电源指示灯和CPU相连，如此它可以指示一个低功率的“备用”状态。更新的模型配备有USB，既能控制监视器设置，也能提供键盘和鼠标的接口。添加LED通信能够提供从电源LED到主机的完全数据回路，这允许添加信息采集器或类似设备用作键。这能替代登陆到电脑的密码，或用作电子商务密码认证设备。类似的技术可以用键盘指示灯来实现。2. 用户可以通过信息采集器从电源LED灯复制她或她的自动洗衣机的全部诊断信息，并反馈到电脑，上传到网上服务点。在洗衣机上，没有特殊显示器或连接器是必须的，也没必要用数据连接到电脑。3. 和熟人之间的电话号码和电子商业名片的交换可以通过两个手机来实现。把屏幕紧挨，之后背景LED灯会交换有关数据4. 可编程电子门铃也许需要它的铃声随季节变化而变换。这可以通过组成或下载一首新铃声到电脑，然后用信息采集器传给门铃来简单实现。不需要移动和卸下门铃（伴随着装线困难）或用昂贵的无线数据连接。5. LED通信可以用作手机电子支付。消费者可以用用户接口和手机的无线数据连接来建立和隐含发的电子支付交易。然后他们可以把手机LED通信设备的电源灯对准自动贩卖机，完成交易。LED的方向性和短范围性是一种优势，因为给用户一个明确自然的指示。6. 用LED通信的廉价玩具可以互相通信来同步行动或在相关玩具中提供紧急行动。他们也能和家用电脑通信来程序交互运行或下载新功能。9. 未来工作LED通信设备的通信范围现在最好的只有几厘米。数据率是每个方向250位/秒。这两个值是反比关系：改变基层协议来获得更长的积分时间可以获得长范围，然而降低信噪比和限制了LED间的最大距离。我们现在正在逐步完善LED通信设备的硬件和软件，使得可以在更大的范围内操作（超过1米），降低数据率，和在最大范围内实现以超过1000位/秒操作。操作环境能够被系统探测到，如此数据率可以根据条件自动上升或下降。LED通信设备将会广泛利用。这将需要几层通信协议的标准进程和光学特性等等。参考文献：1. Mims, Forrest M.,III, Siliconnections: Coming of Age in the Electronic Era, McGraw-Hill, New York, NY, 19862. Mims, Forrest M., III, LED Circuits and Projects, Howard W. Sams and Co., Inc.,New York, NY, pp. 60-61, 76-77, 122-1233. Graeme, Jerald, Photodiode Amplifiers: Op-amp Solutions, McGraw-Hill, 1996, pp.4-7.4. Dietz, P.and Yerazunis, W., Real-Time Audio Buffering for Telephone Applications, in Proceedings of UIST 2001, Orlando, FL, Nov. 11-14, 2001, pp. 193-45. Mobil Speedpass is an RFID payment system. Information is available at 6. Schneier, B., Applied Cryptography, second edition, John Wiley and Sons, New York, NY, 1996, pp. 101-1117. Infrared Data Associates (IrDA) information and specifications can be obtained at the associations web site: 8. Bluetooth information and specifications can be obtained from the Bluetooth SIG, Inc. website: 9. Ullmer, B., Ishii, H. and Glas, D. mediaBlocks: Physical Containers, Transports, and Controls for Online Media, in Computer Graphics Proceedings (SIGGRAPH98), July 19-24, 1998MITSUBISHI ELECTRIC RESEARCH LABORATORIESVery Low-Cost Sensing and CommunicationUsing Bidirectional LEDsPaul Dietz, William Yerazunis, Darren LeighTR2003-35July 2003AbstractA novel microprocessor interface circuit is described which can alternately emit anddetect light using only an LED, two digital I/O pins and a single current limiting re-sistor. This technique is first applied to create a smart illumination system that uses asingle LED as both light source and sensor. We then present several devices that usean LED as a generic wireless serial data port. An important implication of this work isthat every LED connected to a microprocessor can be thought of as a wireless two-waycommunication port. We present this technology as a solution to the “last centimeterproblem”, because it permits disparate devices to communicate with each other simplyand cheaply with minimal design modification.To appear in UbiComp 2003, Seattle, Washington, October 12-15, 2003This work may not be copied or reproduced in whole or in part for any commercial purpose. Permission to copy inwhole or in part without payment of fee is granted for nonprofit educational and research purposes provided that all suchwhole or partial copies include the following: a notice that such copying is by permission of Mitsubishi Electric ResearchLaboratories, Inc.; an acknowledgment of the authors and individual contributions to the work; and all applicable portionsof the copyright notice. Copying, reproduction, or republishing for any other purpose shall require a license with paymentof fee to Mitsubishi Electric Research Laboratories, Inc. All rights reserved.Copyright c ? Mitsubishi Electric Research Laboratories, Inc., 2003201 Broadway, Cambridge, Massachusetts 021391. Version submitted to Ubicomp 2003.Very Low-Cost Sensing and CommunicationUsing Bidirectional LEDsPaul Dietz, William Yerazunis, and Darren LeighMitsubishi Electric Research Laboratories201 BroadwayCambridge, Massachusetts 02139 USAdietz,wsy,leighAbstract. A novel microprocessor interface circuit is described whichcan alternately emit and detect light using only an LED, two digitalI/O pins and a single current limiting resistor. This technique is firstapplied to create a smart illumination system that uses a single LED asboth light source and sensor. We then present several devices that usean LED as a generic wireless serial data port. An important implicationof this work is that every LED connected to a microprocessor can bethought of as a wireless two-way communication port. We present thistechnology as a solution to the “last centimeter problem”, because itpermits disparate devices to communicate with each other simply andcheaply with minimal design modification.1IntroductionLight Emitting Diodes, or LEDs, are one of the most common types of inter-face components. Their diverse applications include numeric displays, flashlights,liquid crystal backlights, vehicle brake lights, traffic signals and the ubiquitouspower-on indicator light.Because LEDs are so commonly used as light emitters it is easy to forget thatthey are fundamentally photodiodes, and as such, are light detectors as well.Although LEDs are not optimized for light detection, they are very effective atit. This interchangeability between solid-state light emission and detection waswidely publicized in the 1970s by Forrest W. Mims 12, but has been largelyforgotten by LED users.1.1Ambient Illumination Sensing with LEDsRecently, we have been investigating improvements for infrared remote controlsof the type commonly used with consumer audio/video equipment. An area ofimmediate interest was the pushbutton illumination used on many remote con-trols. To activate the backlight, you must press a button that is nearly impossibleto locate in the dark! We resolved to rectify this situation.Our first solution was to use a capacitive proximity sensor (similar to theone described in 4) to activate the remote control backlight during active han-dling. Unfortunately, turning on the backlight every time the remote is handledsubstantially decreases battery life, not only because the user often holds ontothe remote continuously but also because the remote is sometimes used undergood lighting conditions when the backlight is not needed. While a mode switchcould be added, this would be little better than the original situation.The obvious step was to add a light sensor to turn on the backlight only whenneeded. CdS photocells are inexpensive, but providing an optical path to the cellwould add significant cost and complexity to the mechanical design. Recallingthe photosensitive nature of LEDs, we decided to investigate using the backlightLED itself as the light detector. We developed a simple circuit for this purposethat requires one additional microcontroller I/O pin, but no other additionalcomponents compared to a traditional LED driver.The success of the simple LED emitter/receiver circuit inspired us to considerother applications. For example, by quickly switching between the forward-biased(light-emitting) and back-biased (light-sensing) modes, it is possible to buildan LED-based light source that appears to be constantly on, but is in factperiodically measuring the ambient lighting level and using this information toautomatically adjust the brightness level of the LED. Our demonstration device,shown in Figure 6, has a capacitance sensor to determine that the device is beingmanipulated and an LED sensor/emitter to provide the backlight function.1.2LEDComm: Bidirectional LED CommunicationWhile the measurement of ambient light levels has many applications, a moreintriguing use of this technology is to transmit data back and forth betweenLEDs pointed at each other. We call this “LEDComm”. We have developedsimple prototypes that allow two-way serial data communication between LEDsover a distance of several centimeters.One possible application of LEDComm is to replace Radio Frequency Identi-fication (RFID) systems (e.g. 5) for payment authorization and access control.To test this concept, we have created an inexpensive keychain-size device calledan iDropper that can receive, store, and transmit data. Unlike RFID systems,iDroppers support true peer-to-peer communication, allowing new functionalitysuch as directly transferring authority between devices without need of a specialreader device.The implications of LED-based data communication are significant. EveryLED connected to a microprocessor can be thought of as a wireless communica-tion port. Compared with other short-range wireless technologies such as IrDA 7and Bluetooth 8, LEDComm has a far more limited range, and a much slowerdata rate. But LEDComm can be implemented at a fraction of the cost, andin many cases, may even be free. This is because LEDComm is essentially asoftware interface technique using existing hardware with minimal modification.LEDComm allows us to implement communication functions in places wheretraditional techniques are too expensive. The power light on many consumerappliances can now become a maintenance port for reading service informationor uploading new firmware. Cell phones can transfer contact information toother phones by holding their displays next to each other. For automobiles, thestandard expensive service connector can be bypassed, and all data transferredthrough the “Check Engine” light. (An automobile owner could even use aniDropper to capture the cars fault log and transmit it to the service centerbefore a service appointment, insuring that the proper tools and spare partswill be immediately available when the vehicle is brought in.) There are manypossible applications.In the following sections, we describe the basic bidirectional LED micropro-cessor interface circuit and its use in the smart backlight. We then give a fulldescription of LEDComm, iDroppers and various applications.2The Bidirectional LED InterfaceLight emitting diodes emit light in a fairly narrow frequency band when a smallcurrent is applied in the correct direction. Because the current-voltage charac-teristic is exponential, it is difficult to control a voltage applied directly acrossan LED accurately enough to attain a desired current; some means must beused to limit the current. In discrete systems, this is typically done by placing aresistor in series as shown in Figure 1. Since most microprocessor I/O pins cansink more current than they can source, the configuration shown in the figure isthe most common way of driving an LED from a microprocessor.VccI/O pinFig.1. Schematic of a typical LED driverThe LED is a photodiode that is sensitive to light at and above the wave-length at which it emits (barring any filtering effects of a colored plastic package).Under reverse bias conditions, a simple model for the LED is a capacitor in par-allel with a current source which models the optically induced photocurrent. (seeFigure 2, 3). It is this photocurrent that we would like to measure.An inexpensive way to make a photodetector out of an LED is to tie theanode to ground and connect the cathode to a CMOS I/O pin driven high.=IphotoFig.2. Reverse-biasing an LED for photosensingThis reverse biases the diode, and charges the capacitance. Next switch the I/Opin to input mode, which allows the photocurrent to discharge the capacitancedown to the digital input threshold. By timing how long this takes, we get ameasurement of the photocurrent and thus the amount of incident light. Thissequence is shown in Figure 3.I/OVccCCINFig.3. LED used as a photosensorThe circuits of Figures 1 and 2 can be combined to create a general bidirec-tional microprocessor interface to an LED as shown in Figure 4. This is identicalto the circuit of Figure 1, except that now the resistor/LED combination isplaced between two I/O pins.Figure 5 shows how the pins are driven for the two modes. Figure 5a showsthe “Emitting” mode where current is driven in the forward direction, lightingthe LED. Figure 5b shows “Reverse Bias” mode, which charges the capacitanceand prepares the system for measurement. The actual measurement is made in“Discharge” mode shown in Figure 5c. Since the current flowing into a CMOSinput is extremely small, the low value current limiting resistor has little impacton the voltage seen at the input pin. As before, we simply time how long ittakes for the photocurrent to discharge the capacitance to the pins digital inputthreshold. The result is a simple circuit that can switch between emitting andreceiving light.Because the circuit changes required to provide this bidirectional communi-cation feature consist of only one additional I/O pin and printed circuit boardtrace (which can be provided at design time for zero additional hardware cost)I/O pinI/O pinCFig.4. Schematic of a bidirectional LED interfaceCVccCVccCINa) Emittingb) Reverse Biasc) DischargeFig.5. Emitting and sensing light with an LEDwe claim that adding this hardware feature to a device is essentially free. Ofcourse, software and CPU runtime are also necessary to make this work.Compare this to the cost of adding IrDA 7 (about $7) or Bluetooth 8(more than $10) to a product. Using even a simple mechanical connector can costseveral dollars because of the required level-shifting and electrostatic discharge(ESD) protection circuitry. Using an existing LED for communication can alsosave manufacturing costs because expensive plastic molds for the housing neednot be altered to accommodate a dedicated infrared transceiver, antenna orphysical connector.3The Smart BacklightThe Smart Backlight is one application of the bidirectional LED circuit. Asnoted previously, the idea of the smart remote control backlight is to turn onthe backlighting before the user has to press a button. Also, to conserve power,we wish to turn on this backlight only when it is actually dark enough to needit.Fig.6. The automatic backlight prototypeTo demonstrate this function we created the prototype shown in Figure 6,with the complete schematic shown in Figure 7. This circuit uses a capacitiveproximity detector to determine handling state. Although the basic capacitancemeasurement circuit is identical to that used in the buffer phone 4, we processthe data to look for active handling (changes in capacitance) rather than simplepresence (increased capacitance). Many users will continue to hold a remote evenwhen they are not actively using it, so the detection of active handling is criticalfor extending battery life. Of course, as soon as the user wishes to actively usethe remote again, any significant motion turns the light back on.VccVccgndGP5GP0GP1GP2GP4GP3PIC12C509proximityelectrodesFig.7. Schematic for the automatic backlightThe smart backlight functions as follows: periodically, the microprocessorwakes from sleep, and measures the capacitance. If no active handling is detected,the processor goes back to sleep. Otherwise, a light measurement is made withthe LED. If the room is dark, it turns on the backlight for at least two seconds.While the backlight is on, it continues to check for active handling. Each timehandling is detected, the backlight timer is reset to stay on for another twoseconds.Since remote controls already contain low-end microprocessors, adding thisfunctionality costs very little. The proximity electrodes can be part of the printedcircuit board, eliminating the need for special tooling. If there are spare I/O pinsavailable, the only additional component is a single, inexpensive capacitor forthe capacitance sensor.One might wonder if the constantly running proximity detector adverselyimpacts battery life. In fact, the circuit draws microwatts of power; the prototyperan continuously for 6 months on a single type CR2032 coin-cell “watch” battery.Remote controls typically use AAA or AA batteries with a storage capacityan order of magnitude higher than the coin-cell, so the power draw would beinsignificant compared to the batteries self-discharge characteristics.4Bidirectional Communication ProtocolsIn our initial experimentation with the smart backlight, we often used LED-based flashlights to test the light detecting circuit. This suggested to us thatLED-to-LED communication was feasible. We constructed a simple test setupusing two identical, generic PIC microcontroller boards with RS-232 interfacesas shown in Figure 8.Fig.8. Bidirectional communication with LEDsThese test boards use a simple protocol for data transfer which allows twounsynchronized devices to phase-lock to each other and exchange pulse-width-modulated data bidirectionally. A basic explanation of the protocol is that thetwo devices take turns flashing their LEDs at each other. A short flash indicatesa 0 or SPACE state, and a long flash indicates a 1 or MARK state.The protocol starts out on powerup with the device performing an idlingcycle, transmitting a 1 millisecond light pulse followed by a 4 millisecond receiveperiod. During the receive period, the device executes 40 light measurements,each one taking 100 microseconds. These light measurements provide only onebit of resolution, i.e. whether the incoming light flux is above or below thedigital I/O pins threshold (nominally about 1.5 volts). With only normal roomlight incident upon the LED there is insufficient photocurrent to discharge thecapacitance below the threshold during the 100 microsecond receive period.The oscilloscope trace in Figure 9 shows the voltage at the LED cathodeduring several light measurements with normal illumination. The vertical scale is1 volt/division and the horizontal is 100 microseconds/division. The capacitanceis initially charged to about 5 volts and then allowed to discharge. Notice that thevoltage never drops below the threshold and so the microcontroller will alwaysread the pin as a 1.Fig.9. A series of light measurements under normal room illuminationFigure 10 is an oscilloscope trace of the same setup, but with the LED beingilluminated by another LED. The capacitance discharges completely during themeasurement period, bringing the I/O pin voltage well below threshold andcausing the pin to read as a 0. The idling cycle continues until at least twomeasurement times in succession indicate “light seen”. At this point, the deviceassumes an incoming pulse of light from a similar device has been detected, andshifts from the idling loop of 1 millisecond ON then 4 milliseconds OFF to aslightly faster synchronizing loop, described next.During the synchronizing loop, the transmitted light pulse is still 1 millisec-ond ON, but followed by a variable number of 100 microsecond light measure-ments. When in the synchronizing loop, the microcontroller will terminate themeasurement set after either 40 are performed, or when the trailing edge of alight pulse is detected. A trailing edge is considered to be found when a pair ofback-to-back measurements both indicate “light seen” followed by ten measure-ments without “light seen”.The execution pattern inside the synchronize loop is therefore composed ofone devices LED on for 1 millisecond, then a 1 millisecond period with bothLEDs off, followed by the other devices LED on for 1 millisecond, and finallyboth LEDs offfor 1 millisecond. Even if the devices have clock frequency errorsof up to 25%, they will still be able to synchronize. The nominal synchronizeloop pulse rate is 250 Hz, with a 25% duty cycle. Figure 11 shows an oscilloscopetrace of two devices in the synchronize loop, firing pulses of light at each other.Fig.10. A series of light measurements under LED illuminationNote that their clocks are completely independent and that all synchronizationis occurring via the LEDs and the base protocol.During communication, data bits are transmitted in asynchronous form. A1 millisecond light pulse indicates a MARK and a 0.5 millisecond light pulseindicates a SPACE. The system normally idles with MARK bits being transmit-ted (the data transfer loop is the same software as the synchronize loop). Duringdata transmission, the format starts with a single SPACE as a start bit, followedby eight bits of data, followed by one MARK as a stop bit. This is similar tothe common 8-N-1 RS-232 format. The top trace of Figure 11 shows the datapulse train of a device that is idling, sending all MARK pulses. The bottom traceshows a device sending both narrow SPACE and wide MARK pulses.To decode the light pulses, the receiving device keeps a count of “light seen”measurements for each execution of the synchronize loop. If seven or fewer light-seen measurements are tallied, a SPACE is recorded; if eight or more are seen,a MARK is recorded. The usual asynchronous deframing (dropping the leadingSPACE start bit and the trailing MARK stop bit) is performed. The resulting8-bit data word is then available to the application-level program. A simplehigher-level protocol allows for error detection and correction.The LEDComm test setup works very well. The underlying protocol trans-mits data at a rate of approximately 250 bits/sec in each direction. The micro-processors buffer the data and connect to a host at 38400 bps. Data transfer isrobust up to a range of approximately three centimeters. Because the LEDs weFig.11. Oscilloscope trace of two devices in synchronized operationused have a fairly narrow beam angle, they permit a pointing error of only about20 degrees.Unlike many other protocols, this system is highly resistant to clock speederrors. Not all of our devices have precise crystal oscillators; some use the inac-curate, internal RC oscillator of the PIC microcontrollers. Even with errors inclock speed of up to 25%, communication is not disrupted. In contrast, errorsover 5% in an RS-232 data link will often cause problems.The cheap, unstable oscillators internal to the PIC devices are actually ad-vantageous in this application. Even if two devices are powered up at the sametime in the exact same phase relationship, they will quickly drift out of syn-chronization enough that a trailing edge will be detected and the devices willsynchronize quickly into alternating flashes. In our devices, this usually happensin under 50 milliseconds. If two LEDComm devices were to both have highlystable timebases (or if both devices derived their clocks from the same source),it would be necessary to insert a jitter source (perhaps based on a hash of thedevice serial number) into the idle loop to assure that the two pulse trains woulddrift out of phase enough for a pulse trailing edge to be detected.An additional feature of this base protocol over a balanced pulse protocol,such as Manchester coding, is that the LED gives a visible indication of idle statevs. synchronized state vs. the data-transfer state. The perceived light brightenswhen ready to transfer data (due to the faster pulse repetition rate) and darkensduring data transfer (due to the short 0.5 millisecond SPACE pulses versus theno-data 1 millisecond MARK pulses).5iDroppersTo act as tangible, portable repositories of information or authorization, we havedesigned and constructed a device that we call an iDropper (for InformationDropper). Like an eyedropper, an iDropper can suck up a small amount of in-formation, hold the information, and then expel the information on demand.Unlike an eyedropper, the iDropper can repeatedly expel the same informationnondestructively.The iDropper is meant to be used in situations where the user wishes totransfer data between devices that, for economic or practical reasons, do nothave a viable user interface. This may be because the data transfer happenstoo infrequently to justify adding a display and keypad to the device, i.e. fordiagnostic and initial setup information. An iDropper can be used to shuttle thedata between the device and another which does have a user interface.The tangible information appliance aspect of the iDropper is similar in effectto mediaBlocks 9. The major difference is that mediaBlocks do not hold anyinformation; they act as tokens for information that is passed along a network.The iDropper itself does hold information and can itself be used as part of thenetwork.The iDropper hardware is composed of a tiny printed circuit board, a singlepushbutton switch (the sole user input), a Microchip PIC16LF628 microcon-troller, an LED (which performs data input, data output, and user output),a 3 volt lithium “coin-cell” battery, a capacitor, and two resistors. There arean additional five solder pads so that extra components can be added for ex-perimentation purposes. The entire assembly is smaller and cheaper than mostcar-alarm keychain remote controls and contains fewer components. A mass pro-duced version should cost less than a dollar more than a similar LED keychainflashlight.The prototype iDroppers are also equipped with an in-circuit programmingconnector which allows us to download code into the microcontroller and tochange the personality of the device. We have also devised a small adapterboard to convert this connector to Microchips standard RJ-11 in-circuit de-bugging module. A pair of iDroppers is shown in Figure 12. The lower one hasthe adapter board attached. The large plastic part visible on the bottom of thelower iDropper is the battery holder the largest component of the device. Wehave left over a centimeter of empty printed circuit board material at the backend of the iDropper so that a hole can be drilled and it can be attached to akeychain.The default iDropper personality program is that of an information eyedrop-per. To suck information into the iDropper, the user presses the button twice andholds it in; the iDropper will then suck in any data stream presented to it andstore it internally. Distinctive flash patterns indicate when the mode has beenentered and when recording has finished. Releasing the button early will abortthe process. To squeeze information out of the iDropper, the user presses andholds the button; the data is then emitted repeatedly, about once a second. ThisFig.12. A pair of iDroppers, one with programming adaptermode appears to the eye like a simple flashlight. This is by design: an iDropperis perfectly useful as a small keychain flashlight.The lithium battery employed in the device will allow over ten hours ofcontinuous use. When an iDropper is not transmitting or receiving data, thePIC microcontroller goes into sleep mode. This lowers the power requirementfor the entire system to below the leakage current of the battery, giving a shelflife of several years.6iDroppers as Intelligent KeysOne of our goals for the iDropper is to use it as an intelligent, programmable key.Although many other technologies are used in intelligent keys (RF and RFID,card-keys, etc.), LEDComm has some distinct advantages. First, it requires nophysical contact so there is no mechanical wear unlike in some card-key systems.Second, unlike RF systems, it is directional and short range so the user hascomplete control over what is being unlocked. This allows a single key to beused for many different locks without the possibility of unlocking the wrongone just because it is nearby. Third, LEDComm is fundamentally bidirectionalallowing the use of challenge/response and encryption protocols which can makethe key very difficult to copy or spoof. Fourth, the visible nature of the LEDallows for some user interface. At the very least, the user can easily tell whetherthe device is operating or if the battery is dead. Fifth, LEDComm readers aremuch easier and less expensive to implement than keycard or RFID readers. Thiscould important in situations where the number of locks is on the same order asthe number of keys.The sixth, and perhaps most interesting, advantage is that LEDComm iscapable of peer-to-peer communication. Any LEDComm device can pass infor-mation or authorization to another LEDComm device (assuming the applicationsoftware allows it). In this case, an iDropper with the standard “suck/squeeze”personality program can learn the unlock code, and pass that to yet more iDrop-pers. This ability to delegate authority is completely unique and not a capabilityof smart cards or RFID tags.To demonstrate this use of the iDropper as an intelligent key, we addeda reed relay and an external power supply to one device, and wired it into thesecurity system that locks and unlocks our sites front door. The iDroppers LEDwas aimed out through the glass windows of the lobby. The iDropper personalityprogram was altered to do nothing until it received the proper (secret) command,and when it received that command, to activate the relay to unlock the door forfive seconds. Figure 13 shows the test setup.We then programmed one iDropper with the correct code, and (as expected)used it to unlock the door. Taking advantage of the LEDComm peer-to-peerability, we passed the unlock code to several other iDroppers which were alsoused to unlock the door.Fig.13. iDroppers being used as a door lock (left) and key (right)7iDroppers, Authentication and SecurityIn some applications, the peer-to-peer ability to transfer information or autho-rization is desirable. In other applications, such as financial and other securetransactions, authentication is as important as transfer, and the uncontrolledpassing of authority must be prevented. An unfortunate side effect of the pro-grammable nature of the iDropper is that there is no guarantee that anotherdevice will respect any “do not forward” data tags that may be inserted by anapplication. Non-transferable authorization and unforgeable proof-of-identity aredifficult problems with many subtleties. A solution for even a highly constrainedscenario would involve hardware, software and cryptographic techniques beyondthe scope of this paper.However, simple cryptography is quite possible and can be used to keepiDropper transactions secure from eavesdropping and spoofing. The microcon-troller used has sufficient power to implement common symmetric cryptographicalgorithms. These require the sender and recipient to share a secret key socommunication between any two devices must be configured in advance. TheiDropper has enough memory to hold many symmetric encryption keys and cantherefore be set up to talk to a number of other devices.Zero-knowledge proofs and public-key (or asymmetric) cryptography 6 wouldenable an iDropper to securely prove its identity and communicate with anydevice that had access to published information; no shared secrets would benecessary. Unfortunately, all published algorithms for these require extensivecalculations and data communication which, although available on modern fastworkstations, are not possible in the extremely limited computational environ-ment of an iDropper. The small lithium battery contains only enough energyto run the 4 MHz processor for a few hours, so a calculation that a worksta-tion could complete in seconds would consume an iDroppers entire battery life.We are currently investigating algorithms for zero-knowledge proofs that requireonly the limited processing we have available.One side effect of the limited data rate and battery lifetime of the iDropper isthat the total amount of data it can communicate in its battery lifetime (about106bytes) could be held in an inexpensive non-volatile memory. This wouldallow the use of one-time pad encryption, either for extreme security or for usewith a very limited microprocessor.8Every LED is a Communications PortAlthough almost every electronic device made today contains a microcontrollerand (theoretically) has sufficient capability to communicate with similar devices,the cost of including a communication link often precludes two devices thatare sitting side-by-side from talking to each other. This is the “last centimeterproblem”.With LEDComm technology, every LED becomes a potential communicationpath. This has broad implications because LEDs are widely used as power-onindicators in microcontroller-based devices. The indicator is usually not wireddirectly to the power supply, but is connected through the microcontroller sothat a minimal user interface (some blinking) is available. With the proper mod-ifications, the indicator can be used to communicate with an iDropper or otherLEDComm-enabled device.The ability to cheaply and easily transfer data between a device with a userinterface and one without will permit designers to add more capability to inex-pensive products. Small, portable products can be carried to the user interfacemachine for interaction there, while larger ones may require that the user carrydata back and forth with an iDropper-like device. Here are some applications ofLEDComm-enabled devices that we
- 温馨提示:
1: 本站所有资源如无特殊说明，都需要本地电脑安装OFFICE2007和PDF阅读器。图纸软件为CAD,CAXA,PROE,UG,SolidWorks等.压缩文件请下载最新的WinRAR软件解压。
2: 本站的文档不包含任何第三方提供的附件图纸等，如果需要附件，请联系上传者。文件的所有权益归上传用户所有。
3.本站RAR压缩包中若带图纸，网页内容里面会有图纸预览，若没有图纸预览就没有图纸。
4. 未经权益所有人同意不得将文件中的内容挪作商业或盈利用途。
5. 人人文库网仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对用户上传分享的文档内容本身不做任何修改或编辑，并不能对任何下载内容负责。
6. 下载文件中如有侵权或不适当内容，请与我们联系，我们立即纠正。
7. 本站不保证下载资源的准确性、安全性和完整性, 同时也不承担用户因使用这些下载资源对自己和他人造成任何形式的伤害或损失。
人人文库网所有资源均是用户自行上传分享，仅供网友学习交流，未经上传用户书面授权，请勿作他用。
|
2:不支持迅雷下载,请使用浏览器下载
3:不支持QQ浏览器下载,请用其他浏览器
4:下载后的文档和图纸-无水印
5:文档经过压缩，下载后原文更清晰
|
川公网安备: 51019002004831号