当你置身于一个以眼镜为显示屏幕，而计算机就在口袋中的世界，你不禁也会希望加速度计就在你指尖可触及之处。那么，无论何时何地，你都能随心所欲地取得你的键盘、鼠标以及空气吉他。
这可不是随意在一家硅谷咖啡厅中听来什么含过量咖啡 因的评论，而是加州大学柏克莱分校教授Kris Pister的研究目标之一──他曾经协助开启物联网(IoT)应用，而今则希望未来能将无线节点封装于一颗芯片(或指尖)之中。
Pister 梦想中的技术节点是在一颗芯片上所打造的完整系统。它不仅包括微处理器与射频组件，同时还整合了传感器、天线、石英晶体，以及甚至是太阳能电源。 Pister在IDTech大会上对与会来宾表示，“你只需要把它放在阳光下，它就会开始与邻近设备互连。”事实上，该装置可能具备将热、动作与RF来源 转变为能量的各种能量采集模块，它还可能利用60GHz短距离无线宽带新技术。
但他表示，要实现这项整合技术还需要三到五年的时间。这项技术还必须推动更多的MEMS进展──这可能会是新技术以参考频率与频率取代石英晶体，以及更进一步的芯片微型化发展等。
柏克莱分校开发出跟大学尺寸一般大小的九轴IMU──Gina，持续朝芯片上节点的目标迈进。
XWcesmc
Pister 在柏克莱分校的传感器与致动器中心(Berkeley Sensor and Actuator Center ；BSAC)积极进行这项研究，目前该中心已经能制造出“跟大学戒尺寸一样大”的原型设备了，他说，“我确定我们将可打造出单芯片的无线节点，而我们也越 来越接近这一目标了。”
尽管未来前景看来一片光明，但Pister认为现阶段的发展道路上仍有乌云笼罩。
石 油和天然气产业促进了工业自动化领域的物联网(IoT)应用。Pister成立的新创公司──Dust Networks(已由凌特科技收购)，还曾经协助这些应用的部署。然而，曾经许多美好的前景预测如今看来都远落于后了。例如，一位市场观察家曾经预测在 2007年以前，IoT的节点数将超越手机装置，从而带来81美元的市场规模。然而，“这并没有发生。我们甚至离先前所预测的数字还有一大段距离呢！”
本文授权编译自EE Times，版权所有，谢绝转载
本文下一页：开放标准的解决方案已经完成
相关阅读：
• 瞄准电池供电型应用，Silicon Labs推全球最节能MCU
• 缺乏共同标准，物联网还只是个传说
• 一个高智能化的物联网枢纽--物联网网关XWcesmc
{pagination}
阻碍物联网成长的问题之一包括缺乏多家供货商共同提供一些可靠且低成本的标准产品。“遗憾的是，大部份的无线产品都起不了什么作用，最后只会让业界像在开倒车，而让大家采用的步调放缓。无线技术的确带来了严峻的挑战，目前在学界与商用领域的人们仍致力于克服种种困难。”
Pister和其他技术专家持续推动基于因特网协议(IP)的一系列标准。新的 IoT 版本标准底层采用IEEE 802.14.4e射频技术，同时并以时间同步的网格网络形式，让射频技术在大多时间中能处于关闭状态，以节省功率。
因特网工程任务小组(IETF)几乎已经完成该标准的最后一部份了。这项称为 6TSCH 或 IPv6 时间同步通道跳频的标准，将可用于提供指定路径，让网状网络中的无线节点可彼此搜寻与配对。
在IETF与IEEE等组织中透过开放标准程序完成这些技术与标准的整合，需要花费好几年的时间。“标准的建立是一段缓慢且繁琐的过程，如果你没有真正强大的求生意志，就不应该贸然投入。”
新兴的标准协议链接IoT与因特网协议(IP) 。
XWcesmc
好消息是工程师们已经打造出 OpenWSN 了，这是采用该标准途径的可用开放源代码建置。“我感觉开放标准的解决方案已经完成了。”
下一步，就等着打造指甲般大小的无线节点了。
本文授权编译自EE Times，版权所有，谢绝转载
编译：Susan Hong
参考英文原文：IoT on a Fingertip Is Researcher's Goal，by Rick Merritt
相关阅读：
• 瞄准电池供电型应用，Silicon Labs推全球最节能MCU
• 缺乏共同标准，物联网还只是个传说
• 一个高智能化的物联网枢纽--物联网网关XWcesmc
{pagination}
IoT on a Fingertip Is Researcher's Goal
In search of the virtual keyboard
Rick Merritt
SANTA CLARA, Calif. — "In a world where the display is on your glasses and the computer is in your pocket, you want your accelerometers on your fingertips. That way, you've got your keyboard, mouse, and air guitar whenever you want them."
This is no overcaffeinated comment heard in a Silicon Valley café. It's the goal of Kris Pister, a Berkeley professor who helped pioneer the Internet of Things and is now trying to pack a node on a chip -- and a fingertip.
The node Pister dreams of is a true system on a chip. It includes not just the microcontroller and radio, but also the sensors, the antenna, the crystals, and even a solar power source. "You just put it in the light, and it starts connecting with its neighbors," Pister told an audience at the IDTech event here. In fact, the device could have a variety of energy-scavaging blocks that tap into thermal, motion, and RF sources. It might even use a relatively high-bandwidth 60 GHz radio over very short ranges.
The technology for the integrated mote is still three to five years off, he said. It requires a laundry list of MEMS advances -- perhaps emerging techniques to replace crystals with reference time and frequency from the Net and more silicon miniaturization.
Berkeley's Gina is a step toward a node on a chip.
The Berkeley Sensor and Actuator Center, where Pister works, currently makes prototype devices "the size of big, ugly college rings," he said. "I am certain we will have single-chip nodes. We are very close."
The future looks bright, but in Pister's view, the present might be characterized as partly cloudy.
The oil and gas industries have led the industrial automation sector in real-world IoT adoption. The startup he founded -- Dust Networks, now part of Linear Technology -- helped enable several such deployments. But many glorious predictions have fallen far short. One market watcher predicted there would be more IoT nodes than cellphones by 2007, creating an $8.1 billion market. "That did not happen. We're not even remotely close to the numbers they predicted."
Several problems have stalled growth, such as a lack of reliable, low-cost, standard-based products from multiple vendors. "Sadly, most wireless products out there are crap, and they end up setting the industry back, slowing down adoption for everyone. Wireless is a really hard challenge that people in the academic and commercial worlds are still very much working on."
Help is on the way. Pister and other technologists have been driving a set of standards based on the widely used Internet Protocol. The IoT version is based on the IEEE 802.14.4e radio that uses a form of mesh networking with time synchronization, letting radios stay off most of the time to save power.
The last piece of the suite of standards is almost done in the Internet Engineering Task Force. That standard, 6TSCH for IPv6 time-synchronized channel hopping, will provide the routing glue to let nodes find one another in a mesh.
It has taken several years for the pieces to come together using the design-by-committee open standard processes at groups like the IETF and IEEE. "Standards are incredibly slow and painful. You should not go into standards if you don’t have a really strong will to live."
The emerging stack of standards marries IoT and IP.
The good news is engineers have already built OpenWSN, a working open-source implementation using the approach. "I feel the open standards solution is done."
Next stop: the fingernail-sized node.
