What if you came up with a new way to leverage a common household device but no one — not even the device manufacturers — believed it was possible? How would you prove you were right?
Apparently, if you were the founders of Cognitive Systems Corp., you’d do whatever it took — up to and including building and manufacturing your own device — to prove it worked.
What was Cognitive Systems’ breakthrough? Company founders Taj Manku and Oleksiy Kravets discovered it is possible to use a common WiFi router, found in almost every home today, to track the movement of people in a house.
Amanda Forsyth, Director of Product Management for Cognitive Systems Corp. , described the solution in a recent interview, saying when you walk through a house “your body is basically like a bag of water, deflecting WiFi signals. [Cognitive’s app] sits on the router … using data analysis AI to figure out that that’s a human walking around.”
This simple concept enables a vast number of applications, from home security to the ability to monitor the health of seniors or other vulnerable people so they are able to remain in their home.
“The solution is quite simple,” she said. “It uses the existing hardware of the average home WiFi router and a small software application, which can be loaded remotely. It leverages AI analysis, and can easily identify and isolate humans in motion, and distinguish them from pets, appliances, and other household items.”
The company has even developed a unique approach to data analytics. Where traditional AI and machine learning applications require an enormous amount of data before they are useful, Cognitive’s founders took a unique approach.
“Our founders Taj and Oleksiy have a philosophy that involves more than just throwing a ton of data straight into an AI algorithm. They think ‘ Let’s first apply statistics, and create a really strong foundation based on statistics and math alone.’ While there’s so much you can get from AI, you can get a lot from just statistics alone. That’s where the company’s focus lies. Add AI on top of that. And so you don’t really need as much data. You don’t have to train the system right from the start.”
Cognitive Systems faced an interesting challenge when it attempted to bring its solution to market. Even though the founders could demonstrate that their system worked, neither the router manufacturers nor the telcos that they wanted as customers were convinced.
Taj and Oleksiy were undaunted. Backed by investors (including BlackBerry’s Mike Lazaritis) who had faith in the Waterloo, Ont. company, Cognitive Systems built its own commercial router. Said Forsyth: “We had to go and build out what was essentially a chip of our own — very expensive — and launch our own consumer product to prove it.”
Cognitive Systems’ first product, Aura, was sold directly to consumers. That successful launch proved to chip and router manufacturers that that they could easily adapt their own products to expose the facilities needed for Cognitive’s system.
But it wasn’t “off to the races” right away. Despite having gone to such lengths to bring its solution to market, Cognitive Systems faced more challenges. Even with third-party chip manufacturers on board, it still had to “convince internet providers that this [was] not science fiction – that it was possible.” Additionally, they had to show that it was possible for their solution to work without interrupting or impacting WiFi performance. “It took a lot of championing to prove this was real and possible,” said Forsyth.
Surprisingly, concerns and challenges around privacy were not a huge obstacle. Forsyth said this is, in part, because of the way the solution works. “There are no cameras or microphones. In terms of privacy, from that angle, it’s really nice. And people really appreciate that. And it’s not like an X-ray, where we can actually pick out, you know, that someone’s doing yoga or anything like that.”
Nevertheless, the company focuses on the security of the data by using sophisticated encryption.
Cognitive Systems now has interest from telcos around the world, who are discovering more and more applications for this unique approach. And while the company was able to launch the solution with minimal data, cloud and AI unlocks more potential from the application.
“For eldercare, the first kind of exciting data analytics piece is sleep,” said Forsyth. “It’s looking back at historical time, figuring out when someone went to bed, and when someone woke up. And I think what’s next is starting to surface — changes in activity, changes in patterns.”
“This is so important for healthcare. The Canadian healthcare system is an easy example. It’s all reactive. For many people, things have to get bad before they visit a doctor.” With Cognitive’s application it’s possible to monitor changes — even changes over time — and use these to determine when a patient might need medical intervention.
As exciting as the new healthcare applications are, Forsyth said the future prospects for the company are amazing. “There are so many potential ways we can go. We’ve demonstrated how it can be used as home security and for elder care. We are currently working on some IoT stuff, like being able to walk into a room and have a light come on. I think where it really gets interesting, though, is when you analyze that data.”
“I personally see it as predictive IoT. So when you look at the smart home today, it’s not smart, it’s connected. Someone has gone in and set up all these different rules. While that’s fine in some cases, it’s not really smart. WiFi motion is super-unique because it creates this layer of context. So now all of a sudden you have a context of where someone is in relative space. You can potentially use AI to do more intelligent things that actually become helpful. I think it’s really exciting where this could go.”