Archive for March, 2018

How do engineers conserve energy in IoT devices? Let me count the ways…

Friday, March 30th, 2018

Embedded devices have traditionally been isolated or confined to a single (often private) network. The Internet of Things (IoT) is different in that it is connected to the internet, may or may not be part of a more extensive system, and is projected as being pervasive someday (i.e., in use everywhere). IoT devices are often low-cost, physically small, require little (if any) human intervention once deployed, and consume low power. Regardless of the industry, there is still the need increase capabilities while lowering cost and power consumption. Devices may be remote or so numerous as to make changing batteries laborious and expensive. Some IoT may need to be disposable, if the environment (e.g., outer space, nuclear contamination sites, etc.) create a situation where changing batteries is impossible.

IoT devices may collect data through physical sensors, process sensor input on a low level (e.g., filter noise, discard unwanted data, etc.), a control scheme, or execute an action (e.g., local indication, alarm, fail-safe actuation, etc.), store data, and transmit data either via wire or wirelessly (i.e., telemetry). These functions consume energy.

F. Samie, L. Bauer, and J. Henkel, “IoT Technologies for Embedded Computing: A Survey”, in CODES+ISSS, 2016.

The layers of IoT computing and the communication between them. (Image: F. Samie, L. Bauer, and J. Henkel, “IoT Technologies for Embedded Computing: A Survey”, in CODES+ISSS, 2016.)

 Methods to reduce consumption might include using a low duty cycle where data sampling is as infrequent as possible without affecting the validity of data interpretation. Circuits can be kept in a sleep state or entirely off until needed. For example, the wireless circuitry may occasionally be woken up to transmit a compressed data file and immediately put to back sleep. Power consumption can be unpredictable since a core can consume different amounts of power at different times, depending upon the application. The CPU can have several modes of power consumption based on a state of readiness to act. IoT that uses several power states can decrease power consumption in the long run but also take progressively longer time for the CPU to act, in the following order: active, idle, sleep, deep sleep, and off. Other means of lowering power consumption include using a processor or SoC that operates at a low voltage. In general, the lower the supply voltage, the lower the power consumption. However, there is a point where supply voltage can be so low that external noise competes with true signals. Operating at a higher frequency can also save power, since more cycles can execute in the same period of real time. Time spent in active mode means energy is consumed at a higher rate. One drawback to CPUs and circuits that operate at higher frequencies is that parasitic capacitances increase, and more heat is generated. Also, the trade-off with operating voltages is that a higher frequency core necessitates a higher minimum voltage level. The challenge of lowest power consumption for computing power means more careful engineering with meticulous attention to detail. Analog sensing of real-world signals from IoT input adds another layer of complexity.

Materials science, chemistry, and physics also play a fundamental role in meeting the challenge of less power consumption while increasing processing power. Technology scaling nodes, with ever-decreasing integrated circuit sizes reaching down into the nanometer scales play a part in lowering energy use without sacrificing performance. Moore’s Law has given industries a wild ride, but it is coming to an end as efficiencies become more complicated to implement. The architecture of the core and process technology make a difference. Choosing the most power efficient architecture for a given IoT application is a challenge in trade-offs as architectures offer different features. A newer process technology, Fully-Depleted Silicon-On-Insulator (FD-SOI), can optimize leakage in both active and standby modes, allowing one to adjust optimization for either power or performance dynamically, as needed. Engineers might apply newer, more power conscious technologies in logic, memory, or adjoining peripherals and interfaces, with the added challenge of adopting a new technology with new, perhaps as-yet unavailable drivers or unsupportive operating systems.

Software can also make an impact. At the lowest layer on the stack, an Instruction Set Architecture (ISA) can affect power consumption, since some instructions can take more cycles to execute than may actually be needed for a given action or command. Such factors may seem like splitting hairs, but as Moore’s Law compresses, engineers find themselves collecting the bits and pieces of power savings for a decreasing power consumption budget to meet industry/market demand. Besides compressing files for transmission so that the wireless circuit is active for the least amount of time, software can be employed to optimize active modes by carefully managing or scheduling tasks. “Race-to-idle” is a term that implies that execution should be sped up as much as possible so that execution (active) time is reduced (assuming that operating voltage is not adjustable.)

One workaround is to trickle-charge through energy harvesting, via solar panels, RF wireless energy harvesting, or converting rejected heat from a nearby device to energy.

Doubtless you can think of yet another tip or trick to conserve energy, but the above illustrates just how much is on the menu for lowering power consumption. The real challenge is to select the best ones for your application and accurately judge the design trade-offs, since tactics in one area can affect your strategy in another.

   

Can Amazon and Apple Improve Healthcare?

Thursday, March 8th, 2018
The 2016 annual U.S. healthcare spend was $3.3 trillion dollars. Of this, 32% was spent on hospital care , 20% on physician/clinical services, 10% on prescription drugs, and the remainder went to other professional services including dental, nursing home, and other areas. The details can be found in a PDF from CMS.gov, a U.S. Centers for Medicare & Medicaid Services federal government website. Not only are residents of the U.S. getting fed up with out-of-control prices, so are the companies that provide healthcare benefits.

According to Reuters, Amazon, JP Morgan, and Berkshire Hathaway plan to form a non-profit company with the purpose of driving down healthcare costs for more than a half-million collective employees. Their objective includes a focus on technology for “simplified, high-quality and transparent healthcare ”that is free from profit-making incentives and constraints.” Apple is opening primary care clinics for its employees and dependents called AC Wellness.

I just finished reading the book An American Sickness by Elisabeth Rosenthal, that details the U.S. healthcare system with a look at how for-profit health care and drug companies are taking advantage of a very broken system. The book discusses how and why pricing can vary wildly for the same procedure or drug, and real-world strategies for avoiding being taken advantage of. It makes sense that people like Warren Buffet, Tim Cook, and Jeff Bezos are taking action against a system that does not reveal pricing and makes “shopping around” for an elective  procedure impossible. Surprises are the backbone of U.S. hospital billing practices. The U.S. healthcare system is dysfunctional. In an emergency room we are faced with a dire situation and we can sign away our rights out of desperation and fear.

It sparks a good deal of hope to know that at least some companies are taking action. In a system where a surgeon can legally bill someone $50,000 for three stitches, it makes sense to vote with your wallet, but in the U.S. Healthcare system we are not told costs ahead of time and are often not in a position to argue if life-or-death is the outcome of a hesitation to choose.

Can these companies do health care better? Is capitalism finally forcing a solution to the problem? Cutting expenses and giving people a better healthcare experience seems to be the objective. Based on Apple’s AC Wellness site, it seems like they are looking for people who can design lifestyle plans for Apple employees, manage population health, and integrate clinical practices and technology “that drives patient engagement.”

Last year Apple CEO Tim Cook was wearing an Apple Watch with glucose measuring capability. If pre-diabetics have an incentive to eat so that their blood sugar levels don’t soar, they might be able to manage their diet and exercise so that diabetes (type-2) can be avoided. The Dexcom G5® Continuous Glucose Monitoring (CGM) System already exists as an FDA-approved system that does not require the user to prick a finger to test blood sugar. Instead, a small patch with a hair-like protrusion is affixed so that blood sugar monitoring can be reported every 5 minutes to a smartphone or smartwatch.

Continuous Glucose Monitoring FDA Approved

The Dexcom G5 Mobile Continuous Glucose Monitoring (CGM) System transmits data on blood sugar levels from a patch with a sub-dermal hair-like needle  to a smartphone or smartwatch. (Image: Dexcom.com)

Apple’s vision may extend beyond financial savings for employees. Last December, Apple and Stanford Medicine announced that they had teamed up to use the Apple Watch in a study for detecting atrial fibrillation in wearers. According to the news release from Apple, “As part of the study, if an irregular heart rhythm is identified, participants will receive a notification on their Apple Watch and iPhone, a free  consultation with a study doctor and an electrocardiogram (ECG) patch for additional monitoring.”

One might be able to avoid the dysfunctional U.S. health care system by maintaining good health through regular feedback using technology, but it’s doubtful that any one of us wearing a watch that gives us feedback on our health stats would actually heed the warnings all of, if not most of the time. Being human, we buy a treadmill and use it to hang clothing on. We buy a gym membership with aspirations to go often but then we never find the time. Genetics plays a large part in the good health sweepstakes, too. We attach aspirations to technology that as humans, not all of us can socially adjust to for maximum benefit. But even the unintended by-products of technology might bring change that makes us better and improves the world we live in. I hope so.