Assuming the human eye could see them, what would the wireless signals that we are surrounded by in our homes, businesses and cities look like? Nothing like the streams of ascii characters depicted in The Matrix, sadly. The reality is more like a sponge, says Cognitive Systems co-founder Taj Manku.
“If you actually see what RF looks like — like, for example, let’s say our eyes could see it it looks very odd. It’s almost like looking into a foam, like looking into a sponge. It’s hard to explain,” he says.
The Waterloo, Canada-based startup is using algorithms to visualize RF fields in real-time and interpret what any distortions and scattered disturbances within these fields are — such as, for example, a disturbance denoting that a person or pet is moving around a room. Or a door or window is being opened.
Given we are increasingly blanketed in RF fields, the idea is to make use of these signals for more than sending messages and looking at pictures of cute cats. Specifically for a whole host of applications powered by a device that can interpret physical goings on in an RF coverage area. The startup’s wireless signal reader device is called Amera.
“We’re building a platform that enables you to use wireless signals in just more than communications. That’s basically the core of the platform,” says Manku. “With that platform we run a bunch of applications depending on the market that we’re going after.”
While existing motion detector technologies can be pretty cheap, the idea is for the Amera platform to replace multiple security-focused applications — creating a value proposition by consolidating different features onto a single, scalable platform.
“We are positioning it as we are replacing with motion and all the contacts in your house. So if you add up the cost of all of that, and if you then take our unit we’re actually cheaper,” says Manku.
He also argues there are privacy benefits from using wireless signals as a detection medium, since — unlike a security camera — there’s no risk of anyone being able to hack into a live feed of your interior. Amera also avoids the scenario of members of the family accidentally or intentionally spying on each other. Albeit the tech can be used to track who is in the house when, by their device IDs. So, for instance, it could be used by parents to keep tabs on kids’ comings and goings.
“You can do video surveillance where you’re actually looking, or you can do our motion detection which sort of localizes. Obviously… most people wouldn’t want to stick a camera in their bedroom. But this particular unit, because it’s not intrusive, is a little bit better — it doesn’t reveal a lot,” says Manku.
It will be able to see motion, and the motion is determined by the distortion that happens in the wireless field.
Cognitive Systems, whose founding team includes chip designers, wireless networking firmware, DSP and wireless standards expertise, has developed its own chipset to power the platform. The chip, called R10, contains four wireless receivers and dual multi-vector processors, along with five custom CPU cores. It’s designed to be a more cost effective alternative to hardware the startup says would normally cost tens of thousands of dollars.
“If you look at the depth of technology that’s required — we built a unit before without building our own chip and it cost around $10,000 to $20,000 to build it, with just buying components,” he says. “If we wanted to do everything in real-time it would be as expensive as $100,000. So the reason why we architected the chip in such a way, and we architected the chip inside the processor in such a way that enables us to run all these applications in real-time but able to deal with all the different wireless standards that are currently out there, and all the different wireless bands. And all the frequencies out there.”
How accurate is the motion detecting tech? That depends on factors such as how many wireless signal emitting devices a user might have in their home, says Manku. Larger homes may need several of its hardware units to be adequately covered, but they also offer cheaper units for expanding coverage. So only one Amera hub would be needed per installation.
“You can definitely detect differences of the volume of something and how it’s sort of moving, and also with regards to where it is approximately in your house or on your property,” he adds. “We actually will measure all the scattered signals — that are scattered off walls, off everything — and as you walk you change that scattering profile. And we’ll also look at the phase distortion you’ll introduce.
“For example somebody’s walking around they actually change the phase of the RF. Which is quite sensitive… So we have a number of algorithms, and so depending on what’s actually happening the weighting of each algorithm is sort of adjusting certain parameters.”
The startup is uncloaking from stealth today, after some 18 months of tech development, to discuss its first product and its initial market focus: security-centric applications, such as motion sensing and detecting hacking or network spoofing attempts. It’s raised an undisclosed amount of funding — saying only that it’s in the region of tens of millions of dollars — from investors including Quantum Valley Investments, the investment fund led by BlackBerry founder Mike Lazaridis.
“We’re building a security type device for home and businesses and enterprise. Basically what it does — it does a number of things — one of the applications is it localizes motion. So it will be able to see motion, and the motion is determined by the distortion that happens in the wireless field,” explains Manku. “And then we have other applications like device ID, what devices are there, what the flow of devices are, how they’re moving around. And that just falls into the category around security.”
“What we’re doing is we’re talking about a platform that is scalable,” he adds. “What we’re saying is look you put this in your home or your business and it enables you to do motion [detection] but then other applications will start being added on… So there’s all these applications that then fall onto the unit.”
Manku says it will initially be selling a b2b service to security company partners, who will then resell its hardware packaged with a subscription service marketed at consumers and/or businesses. The kind of notifications a user would see would depend on the customer facing wrapper the security company partner chooses to apply.
To be clear, whichever security firm is proving the end-user service will have access to the encryption keys in order that they can decrypt and interpret the data being generated by the Amera to push relevant notifications out to their users. So anyone buying into Amera’s system needs to feel comfortable with giving potentially two companies visibility on activity within their home or business network. Although Manku says Cognitive Systems isn’t focused on analyzing the end-user data in this initial security market application itself.
“For security, where you’re trying to protect your property, to see what’s happening, we’re not in any way tracking anybody. We don’t de-encrypt anything,” he tells TechCrunch.
Manku says Cognitive Systems is looking to have a service up and running in the first half of next year, and while it says it’s already working with some distribution and security company partners it’s not disclosing names of these third parties at this stage. The initial product launch will be in the North American market. Pricing has not yet been confirmed but there will be a cost for the hardware on top of the subscription service.
The wider vision is for Cognitive Systems to apply its wireless tech in multiple ways — for example it’s also building a cloud processing platform called Myst which it envisages could be used to, for instance, power smart city applications, such as the ability to locate a lost child in a park, say (if the park in question has been kitted out with its devices). Or to generate a spectrum map of an entire region of a city to track people flow — using another bit of chip hardware it makes (a physics engine, called X10).
“There are other markets that we have where the Myst network actually does way more processing [than in the initial security market application],” adds Manku. “We have other markets also like network analysis, understanding what the network is doing. We’re not pushing that too much right now because we’re only about 50 people so we’re very focused… but we do have other markets that we’re going after in the background.”