At the turn of the century, the premise was that GPS was revolutionary, would work everywhere we needed it, and everything else was old hat. Turns out that we need something that works indoors and in critical outdoor applications without a clear view of the sky, like downtown cores, dense forests, and in-and-out of coverage places like mountain valleys. A Canadian team out of Calgary obtained a couple of key patents and founded a business around navigating with sensors when GPS/GNSS was obscured or just plain not available. Just coincidentally, around the same time, cell-phone and tablet manufacturers were adding these same sensors to their devices so users could readily re-orient screens and play motion video games.
I spend winters in Florida, and in the summer I’m in Calgary, Alberta — of late, “summer” can be a somewhat misunderstood term when talking about weather in that part of Canada; many may recall the devastating floods in that region this year, but nevertheless it’s my home for several months of the year. This year while I was there, I visited Calgary-based Trusted Positioning, Inc., in its offices across from the University of Calgary (UofC), and again at the ION convention in Nashville. I recently had the opportunity to catch up with the staff again and to get a progress update on their unique business and positioning technology.
To bring us all up to scratch on who or what is Trusted Positioning, Inc., (TPI) and where they came from, here’s a brief refresher.
TPI has been around since 2009 as a spin-off of geomatics engineering research by its four founders at UofC. Naser El-Sheimy was the prof, and Chris Goodall (now TPI’s CEO) and Zainab Syed were his grad students when MEMS started to become of interest to the group — their focus had previously been largely on tactical grade IMUs and integration with GPS. They put together a report in 2000 on the opportunity that MEMS offered for navigation, and this started them thinking of potential commercial prospects.
So the following year, two proposals were submitted and ultimately accepted by Canadian government support agencies. This eventually provided start-up funding for what was to become TPI. Chris and Zainab earned their Ph.D.s in 2008, and Jacques Gregory joined them from Queens University in 2009. The first two or three years were tough, and El-Sheimy advises if you are not prepared to give up your existing lifestyle throughout the launch period — family, fun, vacations, finances, even sleep — don’t take on starting such a business. In this case, things ultimately worked out for the founders, and TPI is now launched and doing well.
In those days, the premise was that GPS was still revolutionary, would work everywhere we needed it, and everything else was old hat. Turns out as time passed we wanted something that worked indoors and also in a number of critical outdoor applications where there wasn’t exactly a clear view of the sky — like downtown cores, in dense forests, or in-and-out of coverage places like mountain valleys.
The pre-TPI research at UofC led to a couple of key patents that went with the team into the new business, and as the business grew, new in-house patents began to be developed — all around navigating with sensors when GPS/GNSS was obscured or just plain not around. Coincidentally, cell-phone and tablet manufacturers were then adding MEMS inertial components so users could readily reorient screens and play motion video games, so TPI began to use these sensors for inertial aiding or even inertial navigation for handheld personal navigation.
Nowadays, TPI has around 20 employees, has developed more than 20 distinct patents with more in the works, and has been licensing software since 2011. The initial Canadian government (NSERC and NRC) support funding has been replaced by equity investment of more than $2 million and key strategic partners who have signed on as investors. Plus, a strong technology/business-oriented board has been put together. Well-known industry players John Ladd (ex-CEO of NovAtel) and Werner Gartner (ex-CFO of NovAtel) have joined TPI’s board, and several past NovAtel executive team members have also invested a significant portion of the equity raised to date. It has to be a good sign when industry leaders like those invest and believe in the direction TPI is taking.
The latest advisor to lend support is Google’s “Developer Advocate” Don Dodge — a guy who specializes in picking out key technology companies at the right time, invests in them personally, and then helps guide them to greatness.
Before becoming a “Developer Advocate” at Google, Dodge was the director of Business Development for Microsoft’s Emerging Business Team. He was also part of the leadership for technology start-ups Forte Software, AltaVista, Napster, Bowstreet, and Groove Networks. “Indoor location and positioning technology is the next big thing,” says Don Dodge, “and sensors are the foundation of this technology. I’m excited to work with Trusted Positioning, the market leader in using sensors for indoor location.”
TPI doesn’t only use MEMS inertial sensors (accelerometers and gyroscopes) in phones; it also uses magnetometers, barometers, and available Wi-Fi networks and their associated location databases, GNSS, vehicle speed sensors, user updates, and camera inputs.
As its brochure says, “Sensor solution is always on when moving and provides a consistent accuracy output to seamlessly integrate with all available updates.”
The problem with Wi-Fi is that the databases don’t stay totally reliable — so TPI solves this problem by also collecting data using other integrated sensors for positioning, which can then be used to update the very same Wi-Fi location data. This is one of the market areas that TPI believes it can access, since Wi-Fi positioning is becoming a more common navigation source. TPI would say Wi-Fi should be considered as only part of the solution, as it needs help from other sensors to work well.
The Trusted Portable Navigator (T-PN) navigates while people walk or drive and use their cell phones in any orientation, anywhere and everywhere — including malls, airports or subways. T-PN combines the use of existing smartphone motion sensors with wireless updates (such as Wi-Fi and GNSS) for a complete solution with no extra hardware or infrastructure.
Over the last three years, TPI has developed an entire library of typical profiles for how people move and carry their cell phones. Algorithms detect particular movement profiles and then use appropriate filter adjustments to maintain or improve accuracy when in locations such as urban centers.
T-PN software has been released by TPI this year for integration in any mobile phone, tablet, or PC operating systems, with a view to capturing expanding mobile market applications, such as mobile advertising, indoor E-911, augmented reality, and fitness/recreation. Pedestrian navigation, navigation in parking garages, monitoring the location of devices in store displays, and assisting store visitors to find what they are looking for — all these potential applications are opportunities for TPI solutions. Therefore, TPI has so far chosen to market to mobile OEMs, MEMS and semiconductor manufacturers who can embed TPI software solutions in phones or in MEMS devices or components that go into these phones.
New technology areas that TPI is working on include wearables and using cameras as navigation sensors.
Now that a number of devices such as phones, watches, tablets, and (Google-like) glasses come with Bluetooth tethering, their movement can all be integrated to improve the navigation solution for people on the move.
TPI estimates that its sensor drift is approximately 4-8 percent of distance traveled when operating without any wireless updates. Chris Goodall calculates that adding multiple devices could improve overall accuracy of the navigation solution by up to threefold.
And how the heck do you use a camera as a navigation sensor without a massive visual database? Simple — just focus on a stationary object and calculate the turn rate of the camera/user. Not so easy, really, as continuously detecting stationary objects as the user moves sounds quite complex. How do you differentiate between objects moving and the camera/user moving? “Feature flow” over multiple images is apparently the answer in deriving velocity and turning rate. We’ll have to see when and how TPI will solve this problem and field a solution — but I suspect it may be very soon as TPI is apparently providing sneak-peak, hands-on demos at a number of upcoming trade shows this year.
Lots of companies are working on solutions to the indoor navigation problem, but as Goodall indicated, after first discussing things with TPI and then going off to try to do it themselves, people tend to come back to TPI. Its not as easy as it sounds, and it takes time and lots of trial and error to get anything that works, then making something that works reliably under all conditions is even harder. So TPI is now at the stage, with solutions that work well and work very reliably, that the company is are launching on consumer mobile phones and anticipate larger, mainstream deployments in 2014-2015. Look out for phones with TPI software in 2014 — and there is a rumor that the company may also make its software available to applications developers.
We’ll keep in touch with TPI and let you know from time to time as the company makes further inroads into this new market segment.