An Inside Fix
By Mark Frauenfelder, Wed Jun 01 08:15:00 GMT 2005

GPS doesn't work in buildings or urban canyons. A Silicon Valley startup believes tuning into TV is the answer.


The global positioning system (GPS) is great for providing real-time, 3-D, all-weather vehicle navigation. It works all over the world, even in desolate areas bereft of technological infrastructure. But when the 24 Navstar satellites comprising the GPS went online on June 26th, 1993, few people imagined that, more than 10 years later, people would want to use the system to help them find their parked car or the nearest Starbucks.

GPS was never designed with the purpose of locking onto the position of mobile handsets inside buildings or urban canyons. And since 80% of mobile calls are made inside buildings or cars, GPS in mobile handsets isn't practical.

Nobody knows this better than James Spilker, the lead architect of GPS. In 2000 Spilker and an engineering professor named Matthew Rabinowtiz founded a company called Rosum, which aims to bring GPS-like capabilities to locations that would normally block satellite signals. Rosum CEO Skip Speaks is the former CEO of Kyocera Wireless.

Rosum's technology is based on the triangulation of TV signals, which it believes are well-suited for dense urban environments. The physics make sense: TV broadcast signals, which were designed for indoor use, are about ten thousand times more powerful than signals emanating from GPS satellites. TV signals are also in a lower frequency range than GPS signals, which means they do a better job of penetrating walls. (The next time your inconsiderate neighbor blasts his stereo, notice how you can hear the bass and drums better more than the lead guitar solos. Walls are good low-pass filters, which means they block high frequencies) TV also has a higher bandwidth than GPS, which means it can be used to get a more accurate position fix.

Redwood City, Calif.-based Rosum, calls its technology TV-GPS. It combines signals from GPS satellites and existing commercial TV towers to provide coverage in areas that GPS alone would be unable to serve. For a device to be able to use TV-GPS, it needs a Rosum TV Measurement Module (RTMM), which is a chipset consisting of a digital signal processor and TV tuner. This $40 component (the price is expected to go down when produced in large volume) receives the synchronization signal used by analog TVs to stabilize their vertical hold. Location servers installed in a building use this synchronization signal to determine the distance that the signal travels. Rosum's technology also depends on the installation of several nearby "monitoring units" which have sensitive antennas that monitor the timing and stability of TV signals, and send calibration information to the location servers.

For even greater accuracy, it's possible to install "Pseudo-TV" transmitters around the outside perimeter of a building to provide position fixing within areas as small as a single room. This could be useful for tracking people inside burning buildings. Firefighters could temporarily set down a number of Pseudo-TV transmitters around a burning structure and see where the people trapped inside are (provided they are carrying a compatible device).

Rosum imagines potential applications for its technology to include E911 position fixing and asset and people tracking. Next year, Rosum and its partner Trimble Navigation plan to roll out its first commercial product -- a fleet tracking system.

Of course, Rosum isn't the only company trying to make up for GPS shortcomings in urban situations. There's the Time Difference of Arrival (TDOA) method, which measures the time it takes for a signal to reach different carrier towers to pinpoint location. Rosum claims that its indoor/urban coverage is better than TDOA, and that the infrastructure costs are lower.

And a number of companies, such as Skyhook Wireless (formerly Quarterscope), are using existing Wi-Fi networks to establish the position of mobile devices. Such systems require that a device be able to "see" at least five access points in order to provide a reading with an accuracy of 30 yards or so. The advantages of Wi-Fi based systems are obvious: mainly, the infrastructure -- other people's access points -- is free. And the access points don't need to be open. As long as a hotspot is transmitting a unique ID, a Wi-Fi triangulation system will be able to put it to use.

The disadvantage of Wi-Fi based location is that nobody has control over the infrastructure. Owners of access point can unplug their equipment without warning, which means the database of WiFi access points for a metropolitan area needs frequent updating by either hired or volunteer wardrivers.

Whichever system wins out, the potential demand for the capabilities promised by these technologies is still unknown. The success or failure of Rosum or any other location based services infrastructure provider will rest on the desirability of the applications that will be developed to make use of them.