Do you ever sit in traffic wishing you could conjure up that taxi driver who got you through a gridlock at light speed? If you were in Beijing, you could take advantage of Microsoft Research Asia's prototypical T-Drive, a software program based on GPS data from 33,000 cabs that determines the best route through the city at any given time.
T-Drive combines mobile data with cloud computing to mine the collective knowledge of back alleys and little-used streets that are intuitive paths for many taxi drivers. For more than three months, the researchers timed the trips by combining GPS data from taxi cabs with satellite road maps of Beijing. By breaking the routes up into road segments based on landmarks, they could identify the fastest portions. They also separated weekday and weekend results.
While some skeptics might worry that a less-than-noble taxi driver might take passengers on a longer route, the researchers trusted that, for the most part, drivers in the study were not unscrupulous. And the data seems to prove they were right to think so: in the first demo test, more than 60 percent of T-Drive's routes were faster than routes provided by Google or Bing. On average, T-Drive routes saved 5 minutes for every half hour of driving time.
Though the software does not yet account for real time traffic issues like accidents or detours, joining T-Drive and social networking might eventually let users with mobile devices have a hand in keeping the system up to the minute. This reminds me of the recent DOT statement about a possible mandate to disable all cell phones in cars (see Government Evaluating Cell Phones Disablers in Cars). Could T-Drive be another case against this ban?
Tags: Cars, Computer Simulation, GPS, Maps, Online Community and Social Networking
comments ( )