Short-Range, Low-Power Sensors

What was once the last mile for connectivity is now the last 100 meters.


Over the last 10 years the world has done a remarkably good job of connecting the global wireless world. This is partly because of visionaries, partly because of marketers, and partly just because we can, but mostly because of convenience. We now never need be to be off the wide-area interconnected highway. The last decade has radically changed the way we live. The smartphone and its cousin, the tablet, was the final link to ubiquitous wireless coverage, globally.

Thanks to these smart devices, the immediacy of the Internet, and the now always-on world, we are empowered with the ability to do more, do it better, certainly more quickly than ever before. However, we are finding that to accomplish this we need to be connected to more than just our smartphones and tablets. We need to have everything in our lives online.

So far, that hasn’t happened, mainly because the short-range connected world hasn’t really needed to be connected outside of our own worlds. Today, only about 10% of the last 100 meter devices that will make up the IoT are connected. For the home, we are talking about the often referred to refrigerator, toaster and similar appliances, the alarm system, the media center – things that typically are part of a “home area network.” But the home area network is only one of the small area networks that will have short-range smart devices.

As the IoT evolves, other small cells such as businesses, city centers, malls, theaters, stadiums, event centers, and the like, will connect much of what they have on premise (soda or popcorn machines, vending machines, restaurants, parking garages, ticket kiosks, seat assignments, and a very long list of others). And, there are a very large number of devices that are short-range in all of these various cells.

Only recently, have we developed a pressing need, or desire to put our refrigerators, and everything we have access to while mobile, on the net, morphing the brave new world of the , into the Internet of Everything (IoE). And that will make that last 100 meters—that final frontier of interconnect—a reality.

Murphy makes it happen
As is the case with all forms of technology, Murphy’s Law prevails. The last five years have seen a lot of progress in short range connectivity. Sensor technology, in terms of device footprints, power requirements, costs, and integration has come a long way. Now, virtually every device, from a toothbrush to an airplane can now be connected to the IoT thanks to devices.

One example is the latest ultra-low-power micro controller series from NXP. What makes this device so unique is that it draws such a low (3 μa) current in power down mode, but is still is able to listen, and wake up instantly when told to. The availability of always-on, low-power sensors will enable a much wider portfolio of IoE devices.

Another advancement comes from Rambus in the form of “lensless” sensors. Such sensors are a significant step in the race to miniaturize sensors in that lenses, like LCD screens on laptops, are a significant portion of the cost of the sensor. Eliminating them is a step in the right direction.

“Lensless smart sensors will enable new form factors as well as new usage models that will help to advance IoT,” says Steven Woo, vice president of enterprise solutions technology at Rambus. “Today’s sensor technology is good at capturing things like images, and passing this information to other processing elements that determine whether or not a face is present. In the future, we believe that sensors will integrate both processing and sensing together and be able to extract higher-level meaning that will be passed on to other processing elements. This will enable smaller, less expensive sensing systems that can be deployed across a wide range of products without needing to assemble and integrate multiple separate components.”

It’s a mad mad mad mad wireless world
With licensed cellular radio being the exception, much of the wireless technology development of the last 20 years has spawned a plethora of unrelated platforms (see Figure 1, below). Each platform has its own proprieties, technologies, and protocols. None of them were designed for interconnect with each other or in an IoT world. And few were thought of as low power from the onset.

Fig. 1: The major wireless platforms. Source: DigiKey

It is unlikely that all of these disparate technologies will, all of a sudden, begin to include a universal IoT interface. So what will likely be the solution is some sort of “glue” technology (see Figure 2) for both low- and classic-powered devices.

Fig. 2: The Gatekeeper

Perhaps the most promising glue technologies are gateways. This seem to be an ideal solution for a number of reasons. First of all, they can be designed to handle whatever is required. Second, they can be any number of existing devices (smartphones, for example). And third, they can be software configured to meet various scenarios.

The great enabler to all of this will be with IPV6, and with low-power sensors, 6LoWPAN. Both will be discussed, along with their role in gateways, in detail, in an upcoming article.

Low-power sensor drilldown
Today, just about any parameter desired can be sensed. One thing low-power sensors have enabled is a much broader platform for mobile sensing. “Today, sensors can sense anything—fingerprints, heart rate, blood pressure, perspiration levels, even the cadence of one’s walk or jog,” says Jim Morrison, vice president of the Technical Intelligence group at Chipworks. “The iPhone has up to a dozen sensors, all of which are low-power.”

He notes that all of this is largely due to the advancements in process technologies that have occurred over the last 15 years. One of the major advances that allows such a composition of sensors in things such as smartphones is the ability to work in the digital world. “We now have the ability to put a very slick A/D sensor in devices. It is sensed in analog-land and immediately converted to the digital domain to make it low power. So those sensors, with the advent of such things as MEMS technology and LED technology, makes sensing just about anything possible.”

There are two big considerations for sensors: battery life and size. “There are a lot off applications where one wants to keep the device awake without running down the battery,” says Ross Bannatyne, general manager for the mass-market product line microcontrollers at NXP. “We don’t always have control of when the data is being sent to the sensor. And, power budgets are very critical, A lot of developers walk a fine line on how to add sensors to a network and not blow their power budget, especially since many of these sensors need to maintain a certain battery life because they are remote and there isn’t much opportunity to change or maintain batteries. With IoT sensors, one of the top concerns, other than security, is power.”

Look for technical specification such as 1.8V MOSFETs, with leakage currents of around 10 pA per micron of width. They will feature very low threshold voltage (V^th) of approximately 0.3V. Such specs enable SoC designs that feature very low active and stand-by power consumption, as well very low start-up voltages.

There are a number of companies standing on the edge of such numbers, but this technology is still expensive. Intel is working on a 14 micron platform, which is pretty much the bleeding edge today, and Texas Instruments has a new very-low-power, linear battery charger, and a tiny, fully integrated DC/DC power module, which consumes only 360 nA of quiescent current.

Next up, thingbots
Thingbots are the latest iteration of very common software that runs automated routines or tasks on the Internet. The root reference is bot, which is just a contraction of the term robot, and Thingbots are the bots of the IoT. Legitimately, bots are, generally used to mine and collect data and perform repetitive, mundane, and routine tasks via automated programming.


With the upcoming trillion+ sensors expected to be part of the IoT (IoE), collecting and managing the huge amounts of data that will be generated by sensors would be unmanageable by mere humans so thingbots are a good thing. They will be fundamental in the big data world for collecting the zettabytes of data expected.

Much of the reputation around bots has to do with malicious activity, primarily because no one pays attention to them until they are used for malicious purposes. Bots do have a good side, however.

“Bots of all types will increasingly integrate growing numbers of sensors and make decisions based on the data these sensors collect,” says Woo. “We believe there will be a growing opportunity for sensors that can extract some level of meaning from the data and pass this on to sensor fusion and decision analysis systems. We also believe that this will open up huge opportunities for semiconductor security to secure the valuable data being captured by sensors.”

It will be interesting to see the direction this platform takes as the IoT IoE) evolves.

And, the usual security concerns
Going forward, agencies should develop IoT (IoE) archi¬tectures that would eliminate or minimize threat vectors while simultaneously optimizing performance and functionally. That is a tall order because it goes against the principle of cost containment. But in the long run, it would actually save money and enhance performance because metrics such as uptime would be increased, the cost of threat mitigation decreased, and the costs associated with peripheral security would be lower (manpower, software, threat abatement hours, etc.). Cost reductions also would be seen in manage¬ability, ease of use, resilience, and scalability.

Security should be implemented at all levels. Proprietary sensor protocols that run on or are converted to IP, offer the chance to embed next-generation cryptography. If IoT devices don’t have the resources (storage or computation power for encryption) it can be implemented at points of aggregation, gateways, and access points.


The fog rolls in
The newest entry in the IoT lexicon is the “fog.” Some say it will replace cloud computing, but how that will shake out is yet to be seen. The fog is an emerging, distributed-computing layer that will enable an entirely new class of services and applications to support the hundreds of billions of things in the IoT that will become part of daily lives. Pundits claim that an interplay between cloud and fog is critical because, when it comes to big data management and analytics, unsecured data can have catastrophic consequences.

The fog delivers the ability to collect, aggregate and manipulate data locally, while further analyzing data through enterprise and cloud repositories. Securing connections between IoT devices and access points, gateways and edge routers is key.

There is a brave new interconnect world emerging and it will be chock-full of sensors of every type, shape and variety. Two of the most critical issues that architects will face are security and power. Sensors will present a tantamount problem if there isn’t some sort of common platform that security can be implemented on. More and more architects are looking to hardware for that.

There will have to be some common direction with wireless interconnect. Gateways seem promising, but to have a gateway that will accommodate all platforms and all technologies may be difficult to develop and secure and layer upon layer of glue logic will simply bog down the system. A better alternative would be for the wireless industry to add a common interface that will tie into a security layer that can interface to the IoT (IoE).

When it comes to power, Bannatyne hit the nail right on the head. Sensors will have to use very minimal power in rest or power-down modes, yet instantly respond to input. Whether they are tethered to grid power or not, they will still have to be energy misers. For those that are not tethered to a grid, many will be powered by free energy such as wind, solar or energy harvesting technologies. Batteries will still be a play big in sensors, but new low-power technologies and devices will enable micro-batteries to be the primary, or in some cases, secondary power source.

Finally, will the fog evolve as the next “cloud” to enable a whole new world of services and apps? It’s hard to tell. It is just now being introduced. And, as with any new or evolving space, there is no shortage of prognostications.


Ranger says:

Intel’s at 14nm, not 14 microns…

Leave a Reply

(Note: This name will be displayed publicly)