Wearables Are Distributed Computing’s New Frontier

If you want to see the latest wearables, head to your local electronics or sporting goods store. But if you want to see the latest wearables that are leveraging machine learning at the edge, say hello to Monisha Perkash, Andrew Chang and their colleagues.

Perkash and Chang co-founded Lumo Bodytech, a six-year-old, 30-person Internet of Things (IoT) startup that is not only shaking up the wearables market but pushing the boundaries of distributed intelligence with their designs.

“There is data that works well in the cloud, brute-force computing on macro-level statistics of users and behaviors,” Chang says. “But there are opportunities in the short term because a lot of our products provide real-time feedback using machine learning.”

No one argues the power of the cloud and how it’s transformed application and system design. But as devices get smaller, more powerful and more power efficient, designers want to take advantage of those advances at the edge for performance gains and privacy improvements.
 
There’s data and then there’s actionable data
To understand Lumo Bodytech’s technical approach, you first need an appreciation for the vision behind the company, because that filters all the way down to device design.

“Companies are waking up and realizing they need an IoT strategy. They need to be digital to capture data about their customers and have a more intimate relationship with their customers, develop to their needs better, offer more personalized solutions,” Perkash says.

But data, in and of itself, is meaningless, she adds.

“We’re all drowning in data. What matters is how that data gets turned into actionable feedback that customers find valuable for themselves and their end-users,” she says.

Because the IoT sector is bursting at the seams with competition, companies like Lumo are finding innovative ways to compete—ways that aren’t just about technology.

In Lumo’s case, it’s a three-pronged approach. The company has leveraged Arm-powered technology and integrated SoC design techniques to deliver to the marketplace the Lumo Lift and Lumo Run wearable devices.

The Lumo Lift emerged as a classic “necessity is the mother of invention” situation: Chang was experiencing back problems, and Perkash suggested posture classes as one way to help resolve the problem.  When Andrew took her advice and his back pain decreased significantly, they had their first big “aha moment,” realizing that offering an easy-to-use posture feedback solution at an accessible price point could help improve the lives of millions of people who also suffer from back pain.

So they conceived the idea of a wearable device that would monitor a person’s posture, and help that person learn proper posture by sending vibrational feedback from the device whenever things got out of whack. After successfully launching the Lumo Lift, they then expanded their platform to support the Lumo Run. The Lumo Run is a clip-on wearable that measures and monitors things such as a runner’s cadence, stride, pelvic rotation, drop and bounce to optimize running performance.

“There is so much we don’t know about ourselves and our bodies,” Perkash says. “With the proliferation of connected wearable solutions, we’re starting to get clues.”

There’s clearly a demand for information. Earlier this year, BCC Research reported that the global market for self-monitoring technologies will exceed $20.7 billion this year and triple in size in the next five years.

“It’s getting more exciting as these technologies become more interoperable. They’re starting to create a more holistic picture of people through personalized feedback. That’s the power of IoT,” Perkash said.
 
Opening the platform
That’s one prong of Lumo’s strategy. But another out-of-the-box part of the company’s strategy has been—as of January 2017—to license its platform to companies to allow them to develop their own technologies. This came on the back of the success of the consumer devices.

“We’ve had a lot of inbound interest from large brands,” she says. “They’re keen to leverage the power of IoT technologies, but often creating tech-enabled products and incorporating IoT is not within their core competence.”

The Lumo Motion Science Platform is a wisdom-of-crowds approach that allows companies to drive Lumo technology into areas such as clothing, elder care, jewelry and more applications.
They’ve signed with PUMA to introduce a new product powered by Lumo’s AI MotionScience Platform and are negotiating with others Perkash says.

The third prong of the company’s value strategy is to turn the crank on distributed intelligence, to put more computing closer to the edge to yield more actionable data faster. This is a near-term Holy Grail for technology companies grappling with data-management issues on the back end and security and privacy concerns on the front end.
 
Small, powerful and always learning
Sensors inside the Lift and Run self-calibrate as you use them to personalize the experience and normalize to the individual.

Lumo engineers designed in an Arm Cortex-M4 Bluetooth Low Energy (BLE) SoC using 512kB flash and 64kB of RAM. Chang said that footprint was a consideration in the technology selection, but the biggest driver behind the selection was the operating, BLE and standby power consumption numbers.

“It’s different for each person,” Chang says. “We have to predict the best posture for each user. We have to be constantly adapting to the user and once the device is on and calibrated we have to auto calibrate because wearables move around so much during the day and we have to control for that.”

This is a glimpse at the future of IoT systems and wearables device design that’s just starting to come into focus at companies like Lumo.

What else does the near-term future hold? Lots, say Perkash and Chang.

Chang likes to say the body is innovation’s next frontier. His wish list includes more-integrated SoC with more sensors fusing more data together faster.

Asked what he’d like to see from the ecosystem to push IoT and wearables to a new plane, Chang responds almost before the question is finished: Wireless charging.

“We view charging as an impediment to what we do. We want a wireless future, free of having to charge anything … we’re so close,” he says, emphasizing the words like he’s pinching two fingers together. “That’s when we’ll see IoT really take off.”