December 22, 2015
It seems as though new reports come out every week related to the growth of wearable technology. According to a report published in November 2015, the market for wearable sensors is expected to reach $5.5 billion by 2025. I mention this forecast, since in this article we will be moving away from a device-oriented discussion of the potential for wearables in business in order to consider the sensor technology within our favorite wearable devices–specifically, motion sensor technology. How might enterprises capitalize on motion sensors–cheap, ubiquitous, powerful, and coming soon to a wearable device in your workplace?
At the recent EWTS event in Houston, TX, Toronto-based tech company Kiwi utilized the sensors already present in attendees’ smartphones to track their movements–specifically, their activity levels during networking breaks. From this data, the Kiwi team was able to determine the most active networker throughout the conference, and reward that individual with a prize.
Now imagine the sensors built into your wearable devices. These are closer to your body than your smartphone, and potentially more accurate. Kiwi’s smart gesture recognition platform has the potential to measure some very significant or key motion and activity metrics in a variety of workplaces and enterprise environments based upon the sensors embedded in common enterprise wearables.
So what is gesture or motion recognition?
When we think of wearables tracking our movements, most of us think steps taken, distance traveled, calories burned; but gesture recognition on wearable devices can be exploited to do much more than counting steps–wearable technology can be used to track both how you move and how well you move.
See, wearable devices contain a variety of sensors – video camera, motion sensors, EEG, EMG, ECG – that help track and quantify our day-to-day activities. Gesture recognition essentially allows for the ability to track and qualify all kinds of motions. In a closed environment, different gestures can be tracked and from there outputs such as velocity, power, and torque can be derived. For instance, Kiwi’s smart gesture technology can distinguish a hand-waving motion from other kinds of movements, and furthermore calculate outputs such as power or speed of that hand wave.
The motions are derived from looking at accelerometer and gyroscope data. Accelerometers and gyroscopes can be found in a lot of wearables today, including most smartwatches. Those sensors provide raw data–a seemingly random assortment of numbers generated at a very fast rate. Kiwi’s expertise lies in reading all those numbers and determining what they mean as far as how one moves, or – in the case of enterprise – how individuals physically perform their jobs.
So using an accelerometer and gyroscope, a 3D motion path of an individual gesture or movement can be created. Performing a specific motion amounts to a bunch of data points, which Kiwi then charts to get a graphical representation of that motion. Incoming motions can then be compared against the graphical template.
Ideal form and consistency can also be tracked by creating an optimal or average motion path from the data. The motion paths of different movements can then be compared against that curve to determine to what degree the gesture veers from the ideal or average. This is smart gesture recognition–giving data to processes like a hand wave that are currently untrackable and then allowing for qualification and optimization of that process through data comparison and analysis.
So what kinds of motions can be measured? If it can be imagined it can be tracked, according to Kiwi. Examples of motions that can be identified include device motions (in the case of a smartwatch, “arm up/down,” “arm left/right,” “rotate clockwise”), as well as sport motions (a bicep curl or tennis forehand); and activity states like “start/stop moving,” “turn left/right,” plus sleeping, sitting, standing, walking, running, and various forms of fall detection.
As mentioned, Kiwi can access relevant sensor data provided on many devices, including smartwatches; in addition, small motion sensors can be attached to a CPU then fitted as a unit into a variety of products, such as a hockey stick, a wrench, or a worker’s gloves. So with Kiwi’s platform, many products can be digitized to enable tracking of a desired movement or movements in a specific context–information which can then be analyzed and turned into digestible and actionable outputs or insights.
Read Part II to find out what smart gesture recognition means for the enterprise.
