Fast Company

Why Better 3-D GPS Could Disrupt The Location Business

GPSYour GPS could soon know how high you are.

Rice University scientists have written some smart software that better interprets the signals coming from space that tell a GPS device its altitude, with an accuracy of around a centimer. You may never have noticed it, but even when your car's unit is reporting your position to within a few meters, its guess as to your altitude is very poor--much worse than one centimeter.

GPS systems are all about advanced mathematics, error corrections, and detecting faint radio signals coming from space. The Rice software better corrects for unpredictable errors caused when the radio signals filter through the atmosphere. The upshot is a more reliable and relatively cheap way to make a hyper-accurate GPS system for tracking precise location movements, and because the adjustments are made in software it's not necessarily difficult to implement.

Environmental researchers studying erosion of mountains or the movements of glaciers will be the immediate beneficiaries, but there's also the option of deploying cheap sensor units in dangerous situations to try to predict landslides and other natural disasters.

And when the technology dribbles down to the level of smartphones and stand-alone GPS units, there are implications in the location-based service market--particularly when you look at augmented reality technology. AR uses your position and the direction you're looking in, data gained from GPS and other sensors, to deliver its augmented data layer. If its positional data was accurate down to centimetric scales, it could become even more useful, enabling users to gain data from far more locations and objects. AR gaming could then become a reality, since AR "clues" could be hidden with extraordinarily precise positions. More subtlety, your in-car GPS may never again guess wrongly which road you're on, and--as sometimes happens--gives you erroneous directions as a result. And in fact, there's barely a use of GPS that we currently have that wouldn't benefit in some way from greatly increased accuracy, which will also enable many more emergent uses that we can't imagine now--it's innovative app developers coding for Androids and iPhones that'll dream these uses up.

By listening to the radio data coming from the cluster of GPS satellites in orbit, each of which has a unique identity and radio code, every smartphone and in-car unit can work out where it is by triangulating its position--it's a bit trial and error, and that's why it takes your phone a short while to accurately zero-in. The physics and math don't quite line up to result in super-accurate altitude data, however. And the North-South, East-West position is only ever accurate to a meter or so--military users have access to a slightly higher accuracy, and with very expensive technology you can build a system that's hyper-accurate to less than a centimeter.

[Image: Flickr user avlxyz]

Chat about this news with Kit Eaton on Twitter and Fast Company too.

Add New Comment


  • Andrew

    Dr. Grejner-Brzezinska is from Ohio State University, not Duke.  Also, the first sentence is very misleading since, as you later point out, all GPS receivers really do estimate the height as well.  

    The technique in this article uses double-differencing with a reference station receiver (in fact, multiple reference stations).  These differential techniques have been getting centimeter resolution for a long time (at least 20 years?).  This research is about increasing the distance (and height difference) between the user and reference receivers by using more advanced tropospheric models.  In this case, 10 minutes of GPS measurements are post-processed in order to estimate the position to this accuracy, and the user receiver must remain stationary for this amount of time.  It certainly requires a clear, outdoor environment to work.  It is actually very good research work, but its audience is the geodetic community, not the casual GPS user community.  

    As you can see, this has *nothing* that is applicable to augmented reality, mobile phones, or in-car GPS.  In these applications, the user can be in very complex multipath environments, will not have local differential measurements available, and will not be stationary (to within a cm) for such a long period of time.  Standalone GPS will most likely never be able to achieve what you claim it will.  The only hope of general centimeter resolution for these applications is to augment GPS receiver with additional sensors (inertial, visual, other signals of opportunity) that do better in indoor/urban environments.  If you are looking to do a story on this topic, this is where you would look.

    Please strive to seek out knowledgable contacts that can review your writing.  I still enjoy Fast Company. Thanks.