Slashdot Mirror
Google Maps Directions Adds Real-Time Traffic Estimates
First time accepted submitter constpointertoconst writes "If you use Google Maps to calculate directions, you may now notice (if your route is covered by their traffic data) an 'in current traffic' travel estimate for current route. Some may recall that Google Maps had a similar estimate in the past, but it was removed last year due to poor accuracy."
Yeah, I'll be turning that on.
Deleted
Going to put back swimming and kayaking? And traffic on the sea?
I use Waze on my daily commute, and it allows me to pinpoint the exact location of police speed traps. It's saved me from getting a lot of tickets.
A handful of other users in my area do it as well, it is usually dead-on-accurate.
Unless Google adds in this sort of functionality, it's not the type of thing I would use on a daily basis. Waze is.
then Google users would unwittingly create said traffic conditions. It actually presents an interesting problem in optimization. Suppose that car GPS installations begin to feature the Google Maps traffic estimation. Not only would people unfamiliar with an environment be compelled to use the new feature but even those who are regular auto commuters. Those who wish to save time getting home would all be directed along the same main roads which at the time would report lower traffic. These users would, however, in turn create traffic on those same roads that originally looked so appealing. Unless the units could communicate with eachother to distribute drivers evenly along roots with approximately equal favorable traffic conditions, a system like this would present problems if it were to become sufficiently ubiquitous. It is related to the problem with automated trading systems on the stock market. Many systems at holding companies would observe that a stock has crossed some critical threshold and would begin to buy or sell in vast quantities automatically. The resulting change in value does not contribute to the market's ability to allocate capital, but instead is just a fluke resulting from a Boolean algorithm. All I'm saying is... this traffic business may cause unforeseen issues.
Or at least, that would be a nice way to hack it. Keep everybody off the route you want to take, by having Google tell everyone else to avoid it. Google reports for your route: Bumper to bumper traffic jams, overturned tractor trailers, oil slicks, zombie attacks, etc.
And you'll have an open road.
Schroedinger's Brexit: The UK is both in and out of the EU at the same time!
Where I live, the base traffic is pretty predictable. Certain interchanges will be slow, certain arteries will be slow, but really the random feature are accidents. Some outlying areas have really bad drivers, because there is always an accident. other places not so much. In any case, Google is horrible at providing traffic data. We have a local service that is nearly perfect. There has been times when I have been sitting still on the freeway, and Google has told me that we were moving at posted speeds. No indication of a problem. Frankly I don't use it much anymore.
"She's a scientist and a lesbian. She's not going to let it slide." Orphan Black
I mean, it is nice to know that Google's fastest route is a disaster because freeway traffic that was assumed to move at 50mph is actually doing only 5mph but when is Google or any other map service going to use this information to calculate a route that works well now?
It would even be a nice touch to update the estimates based on a projected time. Right now, I can change the map to show traffic conditions at 9:00am Monday morning but the time estimate is still based on 11:00am Saturday.
Well, I work in traffic software and services area and have spoken with one of the lead developers from Google for the traffic features. The Google Traffic team is located in Zürich, Switzerland by the way. ...)
How it works (and this is not a Google exclusive, TomTom is doing more or less the same, many others as well), at least in general terms:
- Google receives map and/or navigation requests
- If it's from a mobile device and the GPS is activated the requests contain a GPS coordinate of the device
- The location is matched to the underlaying digital road network
- The next location from a subsequent request is matched to the road network as well
- From the difference in location on the road network and in time a vector is created which contains the speed on the road
- Before, during and after this, there are a lot of filters to filter out unwanted data (pedestrians, locations not on the road network,
- After aggregating all data for the whole road network on a individual link basis, one can estimate a level of service and corresponding travel times
It'd be quite stupid NOT to do this with the location data which is contained in the requests anyways. The quality of the traffic state derived by this method is in direct relation to the amount of devices on the road. With the amount of android devices or other devices using Google Maps, they have little competition.
Of course data privacy is a concern, but for the process above, there is no need to track a single source over a long period of time. Some FCD/FPD systems (floating car data / floating phone data) use frequent (every minute) randomly changing IDs. This is enough for getting the vectors required to calculate traffic states.
Try the Waze app. It's a working crowd sourced navigation tool. It actively routes around or at least notifies you of slowdowns ahead, as well as other traffic hazards and police locations.
Where I live (Finland), the cops aren't generally considered to be enemies so I don't want to participate in any application that warns people about their location...
See the "Quest" view at top-right. It gets even better when you zoom all the way in to StreetView.
I've been using this in my Android for over a year now (in Europe).