Half a croissant, on a plate, with a sign in front of it saying '50c'
h a l f b a k e r y
Caution!
Contents may be not!

idea: add, search, annotate, link, view, overview, recent, by name, random

meta: news, help, about, links, report a problem

account: browse anonymously, or get an account and write.

user:
pass:
register,


                           

Route By Example

Leverage human insights in GPS navigation systems
  (+3)
(+3)
  [vote for,
against]

Dashboard navigation systems and online mapping tools generally draw their source data from the same handful of commercial mapping services. For various reasons, this data is error-prone, and sometimes can recommend routes that are really poor choices. (See link for some horror stories).

Meanwhile, a fantastic source of data is going untapped. The local residents of any area often know the best and worst routes to use.

In a simple (single user) implementation of this idea, you would use your dashboard navigation system to RECORD rather than recommend a route. Press a button when you start your journey, then again when you arrive. The system records your source and destination, time of day and elapsed time. Now, when anyone else uses your car, they can choose this pre-recorded route from a menu, rather than relying on computerized guesswork.

But wait, it gets better.

Maybe you're not sure which of three different ways to work is fastest. Select an existing recorded route from the menu, and press the record button again. Another set of data is recorded, and now the computer can compare the two and recommend the route (or sub-segments of it) which worked better based on actual travel time. The more times you repeat this, the smarter it gets.

But wait, it gets better.

Add an upload capability, and these recorded routes can now improve the intelligence of the online mapping systems as well.

krelnik, Jul 14 2004

Wired: "Route of Problem: Bad Online Maps" http://www.wired.co...,1284,54949,00.html
[krelnik, Oct 05 2004]

[link]






       Mapster?   

       +, especially for being able to compare route segments.
egbert, Jul 14 2004
  

       (+) Do GPS navigation systems have traffic condition updates?   

       As more cars are made with GPS mapping systems, this will probably become commonplace, even automatic. Still, it's a good idea, and not yet baked, so (+)
Freefall, Jul 14 2004
  

       In the future, a huge database with records of every journey, including way points and times.
Selecting a starting and destination point will search the database for the various journeys that match, with a listing of best time first. The database could also be searched according to time of day, and day of week.
The data should be anonymous, I suppose.
+
Ling, Jul 14 2004
  

       I think I'd have to find some way to subvert the system. I'd want to reserve the best routes for myself!
phoenix, Jul 14 2004
  

       If everyone who has one of these widgets and the downloaded 'smart route' option drives that same route, won't it get clogged and turn into a 'dumb route'? General idea is excellent, however, and quite doable. Crispy confection.
phlogiston, Jul 14 2004
  

       //won't it get clogged//
Perhaps that could be dealt with, by having each vehicle re-time any route which is being driven again. The new timing is automatically uploaded to add to the aggregate data. As the route degraded, the computer would automatically start picking other routes for you.
  

       Implementing this would require some probabilistic cleverness to avoid "thrashing" of routes. I.e. everyone uses Main St. on Monday, it gets clogged, then everyone re-routes down Elm on Tuesday, making it clogged. Later, rinse, repeat.
krelnik, Jul 14 2004
  

       I have nothing to add beyond stating the obvious: this is an extremely good idea.
calum, Jul 14 2004
  

       If everyone used the same system, clogging the good routes could become a problem. But if you take a page from the street gang play book, many competing services would work, so long as none became dominant.
Laughs Last, Jul 14 2004
  

       I wonder at times if some conventional mapping services don't extrapolate their estimates of driving time based on field reports of average traffic speed over those routes rather than simple arithmetic sums of speed and distance. I have two way to drive to work. One way is about 12 miles and the highest posted speed limit is 45mph with a fair percent of that 40mph and a lesser percent 25-30mph; the other is about 20 miles with the highest posted speed 70mph with a fair percent of that 40mph and a lesser percent 25-30mph. Either way takes me the same (driving time classified) minutes to drive -- so a useful route for a naive driver would have to be derived by a very complex set of constraints. The most simple information that would immensely benefit me as a naive driver would be "Route A: 12 miles. Be prepared to drive 55mph to maintain traffic flow."
dpsyplc, Jul 14 2004
  

       I'm wondering if there may not be far too many variations on each "point A to point B" for the system to be workable. Any significantly lengthy journey may have several equally sensible routes, some differing only in details.
angel, Jul 15 2004
  

       Indeed angel, and programs like AutoRoute Express can calculate the best journey based on average speeds and road types within seconds.
Sounds to me like what we've got here is a combination of a GPS mapping system with the "intelligent" routing of an auto route express type program, + you can give feedback to the RouteFinder so that it's journey times become close to the truth. You could even use a genetic algorithm to allow it to improve other route estimates based on the comparison bwteeen it's calculated route times and the ones entered by you. Sounds like a fantastic invention to me. Croissant+
goff, Jul 15 2004
  
      
[annotate]
  


 

back: main index

business  computer  culture  fashion  food  halfbakery  home  other  product  public  science  sport  vehicle