Internet of Things

« Back to Homepage

Tweet

Powerful SoCs Turbocharge Wearables’ Future

Just a few months after purchase, yesterday’s wearable was likely to be in a dusty drawer. However, new chips promise enhanced connectivity and compelling applications ensuring the latest generation of products stay on the wrist.

Reports on the wearables sector bring mixed messages. On the one hand, analysts talk of a buoyant market in its infancy with enormous potential while on the other detractors say wearables are a solution looking for a problem. According to The Wall Street Journal, for example, some 111.1 million wearable devices will ship in this year adding to the nearly 40 million that Americans alone wore in 2015. On the downside, analyst Endeavour Partners says buyers of wearables “have subsequently abandoned them at higher rates than other mainstream consumer products.” The company goes on to report that 35 percent of American consumers stopped using their wearable within six months. Moreover, of the one in ten American adults who own some form of activity tracker, half of them no longer use it.

It seems that while counting steps and measuring sleep quality can be motivational for someone wanting to get off the couch and get a little more fit, the novelty wears off quickly if the data is not analyzed and used to make suggestions of how to move forward once the user has attained some degree of fitness. The key challenges for the manufacturers are that wearable technology is an emerging sector and is still searching for the ‘killer app’ that will make the devices a “must-have” accessory rather than a novelty.

We have been here before. Early MP3 players clearly had advantages over products like the portable CD players with which they competed. A user could jog without the player skipping for starters. However, the early devices did not catch on because loading music was difficult and slow, and small memory capacity limited the number of tracks. Many Rio MP3 players went the same way as today’s wearables, to the bottom of a dusty drawer. The MP3 player finally realized real promise when Apple offered iPods with significant memories and an integrated system (iTunes) that allowed an organized, easy, and fast download of songs. Now that integration has reached a logical conclusion with the MP3 player itself having merged with the smartphone.

The current crop of commercially-successful wearable products subdivides into fitness bands—with functionality limited to things like tracking activity and monitoring sleep—and smartwatches, devices that do much that fitness bands offer but add many of the functions of a smartphone such as texting and notifications as well as supporting several built-in apps. The split between the two is roughly 50/50.

However, all wearables have smartphone connectivity in common, which is almost exclusively provided by a power-frugal form of Bluetooth wireless connectivity. Such connectivity allows the smartphone to analyze the wearable’s data and present the information using an application. The smartphone also acts as a gateway to the Internet, uploading the user’s data to a Cloud service for further slicing and dicing. The demarcation between the two devices – smartwatch and fitness band - is becoming increasingly blurred as fitness bands gain greater computing power and smartwatches shrink and run longer on their batteries.

Figure 1: Wearables work in tandem with smartphones for app support and Internet connectivity.

Moreover, wearables makers such as market leader Fitbit are not sitting around waiting for that killer app to spring from nowhere. The company is intent on evolving its products to improve their ‘stickiness’ with features such as constant heart rate monitoring and “guided breathing sessions” which claim to calm the user if their heart rate becomes a little erratic. Fitbit and others are also exploring how wearables could assist sufferers of chronic ailments such as Type-2 diabetes. Regardless of the actual merits of guided breathing sessions, such developments demonstrate how wearables are beginning an advance that will likely mirror the progress that took the MP3 player of the mid-90s from a barely-used novelty to an integral part of today’s monolithic shiny black slabs that are the all-powerful Apple iPhone and Samsung Galaxy. Much of that development will be propelled by a new generation of chips that combine powerful yet efficient embedded processors, interoperable wireless connectivity, lots of Flash and RAM, ever more complex algorithms and sophisticated RF protocols (“stacks”) into tiny slices of silicon.

The heart of a wearable

The wearables market should be very thankful to the Bluetooth^® Special Interest Group’s foresight in merging Nokia’s Wibree Alliance into its organization back in 2007. Nokia’s vision was for its handsets to be the center of a Personal Area Network (PAN) of wirelessly connected “peripherals.” (The company wasn’t entirely clear what those peripherals would do although—taking the lead from wirelessly-connected sports watches of the era, it did recognize that wireless sensors such as heart rate monitors would likely provide a precedent.) Bluetooth^® wireless technology had already found its way into mobiles but was too power-greedy to embed into these peripherals. Consequently, Nokia and its alliance partners started work on an “ultra-low power” wireless technology that could run off small batteries while linking peripherals with the mobile.

The drawback of Nokia’s vision was that, apart from itself, phone makers were not sold on the need for yet another radio chip in their handsets in addition to GSM, Wi-Fi, and Bluetooth^®. The Bluetooth^® SIG spotted an opportunity to overcome this objection by building on Nokia’s work with a form of technology that was interoperable with standard Bluetooth – thus making it part of an established international standard and eliminating the need for an extra chip in handsets. And so, after considerable behind-the-scenes engineering effort, Bluetooth^® 4.0 was born in 2010. Bluetooth^® Low Energy (often, but not officially, abbreviated to “BLE”)—the ultra-low power form of the technology—formed a vital part of the specification.

Bluetooth^® Low Energy (BLE) supports a useful one-megabit raw data bandwidth yet can be powered by batteries as small as coin cells. Better yet, mobile operating system vendors such as Apple, Android, and Microsoft have gone out of their way to encourage BLE device connectivity by providing application programming interfaces (APIs) for their various software packages to make life a bit easier for app developers. It is perhaps of little surprise that sales of BLE components have skyrocketed. Analyst ABI Research, for example, reports that the technology will experience a 34 percent compound annual growth rate (CAGR) between 2016 and 2021 to reach a staggering 1.35 billion shipments.

However, by itself, BLE is simply a wireless connectivity technology – albeit an impressive one. It combines a physical layer comprising a 2.4GHz silicon radio with a firmware stack to manage communication. Early chips demanded a separate low-power microprocessor such as Texas Instruments' (TI) MSP432 to supervise the operation of the RF transceiver. It was all a bit complicated and fiddly to get working, more black art than engineering. Everything changed with the introduction of Norwegian firm Nordic Semiconductor's nRF51 Series BLE System-on-Chip (SoC) in 2012. The SoC was the first to integrate an ARM^® Cortex^™-M0 processor, together with a high link budget 2.4GHz radio, and a decent amount of Flash and RAM, power management, ADCs and DACs, and a smattering of I/Os. Such a design allowed wearable developers to avoid two-chip wireless solutions by employing the ARM device for running both the application as well as the wireless connectivity.

Two-chip solutions were ill-suited for space- and power-constrained wearables because such designs demanded the processor run both the application code and the wireless connectivity (a specialized function for which a general-purpose processor was not ideal). Moreover, the developer was forced to work with two mutually-exclusive development environments and then hope that the application code and BLE stack played nicely together (which didn’t happen very often). It was a challenge not beyond RF experts at sports watch makers like Suunto and Garmin but was a “big ask” for an engineer working at a wearables start-up. And that was without mixing in the problems of extra complexity, greater PCB real estate, and higher power consumption that a two-chip solution introduced.

The introduction of the ARM-cored BLE SoC was the real catalyst for the wearables sector as we know it today. The selection of ARM was no accident; the processors were designed from the outset for low-power, mobile applications, and their worldwide popularity ensured a well-supported, familiar development environment with extensive open-source code libraries. In another clever move, Nordic introduced a development environment that allowed the wearables developers to focus on their application code without worrying about the software integration with the BLE stack. That part of the job was taken care of by the factory’s tools and considerably eased the intricacies of RF engineering - allowing the wearables designer to focus on product differentiation.

Nordic was quickly followed to market by Broadcom, now Cypress Semiconductor, Dialog Semiconductor, NXP, and STMicroelectronics, each of which introduced similar ARM^® Cortex^™-M0 or -M3-powered BLE SoCs with 2.4GHz radios, Flash and RAM plus peripherals on a single chip.

Figure 2: Block diagram of Cypress Semiconductors' BCM20732S, one of a current crop of highly-integrated BLE SOCs from several semiconductor vendors. These SoCs integrate ARM processor, 2.4GHz radio, Flash and RAM, power management, ADCs and DACs, and I/Os onto a single silicon slice. (Source: Cypress Semiconductor)

For now, Nordic Semiconductor has stolen another lead with the introduction of its nRF52 Series in 2015. The device sports an even more powerful processor, an ARM® Cortex™-M4F, together with a more sensitive 2.4GHz radio plus increased Flash and RAM memory. The company says that the SoC is “a single-chip solution designed to cope with even the most complex BLE applications.” What’s more, despite the increased capability, the chip consumes about half the power of the nRF51 Series, extending battery life in the end product.

However, this is just the beginning. Smartphone connectivity has cemented BLE as the premier wireless technology for wearables, and the market is expected to expand to 245 million units a year and be worth $25 billion by 2019, according to analyst CCS Insights. That is a large pie of which other silicon vendors will want a slice. Expect to see those vendors react to Nordic’s introduction by introducing ARM^® Cortex^™ M4-powered BLE SoCs soon, and development of even more powerful products to accelerate.

Development and diversification

Armed with the latest generation of Bluetooth^® Smart SoCs, OEMs have been striving to enhance the wearables experience by following a path that has previously been well-trodden by sports watches – expensive devices boasting GPS and beloved by keen amateur and professional athletes alike. Sports watches are designed to crunch speed, heart rate and power data into guidelines that assist in improving performance. However, whereas Garmin, Suunto or Polar’s devices are hardly svelte and demand almost daily recharges, mass-market fitness wearables have to be unobtrusive and run for a week or more without recourse to the charger.

Fitbit’s recent introduction of the Charge 2, for example, illustrates how the company is taking the heart rate data from the device’s built-in monitor and processing the information to offer more precise calorie consumption calculations, how cardio fitness is changing over time (and how the user’s fitness compares with people of a similar age) and notification of those guided breathing sessions for when stress levels rise. While Fitbit is not about to divulge the secrets of its product’s design, an educated guess would suggest that many of the algorithms supporting this additional analysis are run on the BLE SoC’s processor at the heart of the device.

The company and its competitors are also promoting their products to a wider audience, including sufferers of chronic ailments such as diabetes. The current crop of wearables is not capable of directly measuring blood glucose—the key measure of diabetic health and something which is currently notoriously difficult to track reliably with non-invasive sensors—but the band can help motivate people with diabetes to exercise. Daily movement helps to both reduce high levels of blood glucose and mitigate the long-term health effects of the ailment such as cardiovascular disease. For example, researchers found that just 1,400 steps per day decreased diabetes-related deaths by 21 percent.

BLE SoC-powered wearables are currently introducing additional functionality that includes heart rate variability and muscle oxygen levels. The variation in the period between heart beats is an important indicator of heart health. Two individuals with identical heart rates can have a very different underlying heart condition. In healthy people, low variability is a good sign of overtraining and can be used by the wearable to advise the user to back off the running schedule a tad. For those who have had heart surgery, low variability can be an indicator of poor recovery or possibly an impending heart attack, according to Isansys Lifecare, a British health-care company. Muscle oxygen indicates how much oxygen is in the blood where oxygen is used (compared with traditional measurements performed, for example, at a finger where the saturation of oxygen can be much higher). Muscle oxygen is a good indicator of how close a runner is to exhaustion, as well. U.S.-based company Moxy is among the pioneers of the technique.

However, getting out and exercising and improving fitness, while good for everyone, is not enough motivation in itself for the majority to invest in a wearable; OEMs believe that living a “more contented” life in general probably has a wider appeal. Consequently, BLE-powered wearables that target “proactive lifestyle management” are starting to appear on the market. Such products are increasingly blurring the lines between medical- and lifestyle-wearable devices to the point where the U.S. Food and Drug Administration (FDA) has defined a new category for them, "general wellness devices." A typical entrant to this category is Vinaya, whose wellness wearable, ZENTA, will launch in spring 2017.

The device is equipped with an accelerometer, a microphone, biometric sensors, and a haptic engine that gather noise, movement, and biological signals from the user including heart rate, heart rate variability, breathing patterns, electrodermal activity (EDA), skin temperature, pulse transit time, pulse wave velocity, and blood oxygen saturation. ZENTA combines the signals with learning algorithms to build a profile of the user helping them “live a more balanced life.” Whether Vinaya’s claims for ZENTA stand up to scientific scrutiny is for others to debate, but it does nicely demonstrate how the Nordic nRF52832 BLE SoC at the heart of ZENTA can cope with continuous input from multiple sensors, crunch complex algorithms, and seamlessly connect with a smartphone.

Figure 3: Vinaya’s ZENTA wearable is typical of a new generation of devices that combine multiple sensors with complex algorithms to offer users greater value from the product. (Courtesy: Nordic Semiconductor)

Wearables will also have a big future beyond the consumer sector. One example comes from Honeywell, a U.S.-based automation and control company. Honeywell has teamed up with chipmaker Texas Instruments to develop monitoring for firefighters. BLE SoC powered-wearables carried by the firefighter monitor things like heart rate, breathing, movement, and gestures. The wearable then sends the information to a mobile device (similar to a smartphone but more rugged and powered by the Intel^® Quark^™ SE microcontroller) which forwards the data to a cloud via Wi-Fi or cellular network. The idea of combining signals from various sensors is that, unlike one signal, for example, heart rate, a combination of signals gives a clearer picture of whether the firefighter is in distress.

Figure 4: Honeywell and TI are collaborating to extend wearables beyond the consumer sector. In this case, the companies have developed a mobile hub to collate wearable data and forward it to a cloud. (Source: Intel)

Integration is Key

The first generation of wearables was like the first generation of mobile phones – useful in their niche but limited beyond it. However, as BLE SoCs, batteries, firmware, and supporting platforms have improved, so have the products.

However, it is still the case that unless the user is training for triathlons even the most sophisticated of products might not have sufficient appeal to be worn for years. Nonetheless, if the wearable at least proves its worth by offering useful analysis and guidance from monitored data—such that the user begins to appreciate the product category as more than a novelty—next year’s product will offer enough value to encourage a further purchase. According to Forbes, over 50 percent of U.S. iPhone owners upgrade their devices as soon as their provider allows, even though the current handset is perfectly serviceable; why should the wearables sector be any different?

However, the real key to wearables success in the longer term is likely to be integration. Texting, Internet connectivity, an MP3 player, and camera—not to mention the Bluetooth wireless link to peripherals— have added to a handset’s basic functionality to make today’s smartphone indispensable. In the same way, fitness tracking, stress management, assistance with chronic ailment mitigation, and work productivity and safety tools (and probably more) could come together to make wearables a vital consumer product. That said, for the future of wearables is becoming increasingly difficult to predict. According to analyst Endeavour Partners, much of the innovation is coming from the interface between technology and human physiology and biology. Endeavour Partners says this is an area that very few people understand well and it is not obvious what is or is not practical on a mass-market scale.

However, whatever commercial applications wearable developers eventually come up with, the silicon vendors will work hard to have a BLE SoC that is more than capable of coping with the computing and communication resources they demand.

Lynnette Reese is a member of the technical staff at Mouser and holds a B.S. in Electrical Engineering from Louisiana State University. Prior to Mouser, she completed a combined 15 years in technical marketing in embedded hardware and software with Texas Instruments, Freescale, and Cypress Semiconductor. She started her career as an applications engineer at Johnson Controls.