How do electronic traffic times signs work?
April 29, 2008 5:42 PM Subscribe
How do those overhead signs on expressways know how fast traffic is moving in order to calculate a travel time estimate? And when the traffic is light, do the estimated times on those signs reflect the legal speed limit or the current speed of the traffic?
These are the electronic signs that change to reflect traffic times to the next major exit/entrance, such as "I-55 6 mins."
These are the electronic signs that change to reflect traffic times to the next major exit/entrance, such as "I-55 6 mins."
The estimated wait times at the George Washington Bridge into New York are generally comical. If it says 45 minutes, bring food and drink.
posted by hexatron at 5:57 PM on April 29, 2008
posted by hexatron at 5:57 PM on April 29, 2008
Based on dozens or hundreds of personal observations: The sign on I-94 northbound at Kenosha, WI never shows faster than 12 minutes to Hwy 20, whether traffic is moving at about the speed limit or everybody is doing 85. So I'd guess there is some sort of limit programmed in.
posted by indyz at 6:08 PM on April 29, 2008
posted by indyz at 6:08 PM on April 29, 2008
Ok, I just found the sign on Google Maps using Street View (hooray, Internet!) and measured the distance to be about 12 miles between the sign and the exit. The speed limit is 65, so even if you were doing the speed limit you would make it in less time than the sign said, and traffic (at least when I'm on this particular highway) is almost always actually going 75+ mph.
posted by indyz at 6:21 PM on April 29, 2008
posted by indyz at 6:21 PM on April 29, 2008
I'm pretty sure in major metro areas that have toll tags of some sort, the toll tags are used to track some individual cars as they pass by readers spread across the highway (even if the road is a not a toll road).
posted by comwiz at 6:58 PM on April 29, 2008
posted by comwiz at 6:58 PM on April 29, 2008
I used to work in a traffic engineering research lab for the Florida DOT, and we worked with traffic sensors quite a bit, including both the common means and the uncommon.
There are two major ways to detect the speed of traffic. The most common one, at least in Florida, is microloops - basically big inductive loops that are installed just below the surface of the roadway that can measure ferrous metals moving through an electric field. A controller measures "peaks" generated by the cars as they move over them. Usually one is used to count traffic volume, and pairs are used to measure speed ( d = r * t ). Just FYI, the downside to this system is that it must be replaced every time the road is resurfaced. That can get expensive.
The other, less common method is video detection. Cameras watch a section of roadway and perform real-time object detection. They can basically measure where a car's leading or trailing edge is, and calculate the speed from that. Obviously, this is susceptible to things like weather, smoke, obstruction, etc.
[Some of the really weird methods are weight sensing (far, far less common than most people think), audio detection, radar, and a new type of microloop that's installed in a vertical tube near the roadway. Each has their costs and benefits.]
I can't speak authoritatively on the subject of how the speeds of traffic are reflected on VMS/DMSes (that's what the road message signs are called), but the ones here in Atlanta never show a speed that's faster than the speed limit, and estimated times are never faster than 1 mile per minute.
posted by TheNewWazoo at 8:56 PM on April 29, 2008 [1 favorite]
There are two major ways to detect the speed of traffic. The most common one, at least in Florida, is microloops - basically big inductive loops that are installed just below the surface of the roadway that can measure ferrous metals moving through an electric field. A controller measures "peaks" generated by the cars as they move over them. Usually one is used to count traffic volume, and pairs are used to measure speed ( d = r * t ). Just FYI, the downside to this system is that it must be replaced every time the road is resurfaced. That can get expensive.
The other, less common method is video detection. Cameras watch a section of roadway and perform real-time object detection. They can basically measure where a car's leading or trailing edge is, and calculate the speed from that. Obviously, this is susceptible to things like weather, smoke, obstruction, etc.
[Some of the really weird methods are weight sensing (far, far less common than most people think), audio detection, radar, and a new type of microloop that's installed in a vertical tube near the roadway. Each has their costs and benefits.]
I can't speak authoritatively on the subject of how the speeds of traffic are reflected on VMS/DMSes (that's what the road message signs are called), but the ones here in Atlanta never show a speed that's faster than the speed limit, and estimated times are never faster than 1 mile per minute.
posted by TheNewWazoo at 8:56 PM on April 29, 2008 [1 favorite]
Oh, for your future google enjoyment, the topic you're asking about falls under the broad umbrella of ITS - Intelligent Traffic Systems. One of the big technologies in use for controlling those DMS (dynamic message signs) and VMS (variable message sign) systems is NTCIP - an SNMP-based protocol used to query and set traffic control devices.
posted by TheNewWazoo at 9:00 PM on April 29, 2008
posted by TheNewWazoo at 9:00 PM on April 29, 2008
Er, crap. Intelligent Transportation Systems.
posted by TheNewWazoo at 9:01 PM on April 29, 2008
posted by TheNewWazoo at 9:01 PM on April 29, 2008
Cameras watch a section of roadway and perform real-time object detection.
They can also just have a guy watching a set of cameras in a control center, and he can see how far the traffic is backed up. That plus past experience can give a good enough estimate for a road sign.
posted by smackfu at 9:02 PM on April 29, 2008
They can also just have a guy watching a set of cameras in a control center, and he can see how far the traffic is backed up. That plus past experience can give a good enough estimate for a road sign.
posted by smackfu at 9:02 PM on April 29, 2008
To add another data point, the driving times on the sf bay area signs never estimate anything faster than the speed limit when traffic is moving at speed or above. I routinely beat their posted times by a 15% margin; evidently, I speed a lot.
posted by nerdcore at 11:03 PM on April 29, 2008
posted by nerdcore at 11:03 PM on April 29, 2008
smackfu said: They can also just have a guy watching a set of cameras in a control center, and he can see how far the traffic is backed up. That plus past experience can give a good enough estimate for a road sign.
The whole point of having a system like this is to make it as automated as possible. While there is someone monitoring the system for malfunctions or emergencies they are not constantly updating the signs with traffic times.
posted by JJ86 at 6:03 AM on April 30, 2008
The whole point of having a system like this is to make it as automated as possible. While there is someone monitoring the system for malfunctions or emergencies they are not constantly updating the signs with traffic times.
posted by JJ86 at 6:03 AM on April 30, 2008
In the Bay Area they use FasTrak (RFID) transponders, which don't seem to be any more accurate than any of the other methods.
posted by kirkaracha at 6:42 AM on April 30, 2008
posted by kirkaracha at 6:42 AM on April 30, 2008
Here is a snippet from an ITE article about the WISDOT TMS(Traffic Management System) in the Milwaukee area:
WisDOT determines travel times using data collected from loop detectors spaced 400 meters (one-quarter mile) in urban areas..... To calculate travel times, an algorithm averages the speed measured by the loop detectors of all lanes of traffic. The averaged speed is assumed over a pre-defined link and a travel time is calculated using the distance of that link. Each travel time segment comprises multiple links; travel times are added together to determine the segment travel time.
I would assume every DMS (Dynamic Message Sign) in the US uses that method.
posted by JJ86 at 6:44 AM on April 30, 2008
WisDOT determines travel times using data collected from loop detectors spaced 400 meters (one-quarter mile) in urban areas..... To calculate travel times, an algorithm averages the speed measured by the loop detectors of all lanes of traffic. The averaged speed is assumed over a pre-defined link and a travel time is calculated using the distance of that link. Each travel time segment comprises multiple links; travel times are added together to determine the segment travel time.
I would assume every DMS (Dynamic Message Sign) in the US uses that method.
posted by JJ86 at 6:44 AM on April 30, 2008
WSDOT has a detailed explanation of their calculation method here.
posted by crazycanuck at 6:47 AM on April 30, 2008
posted by crazycanuck at 6:47 AM on April 30, 2008
These signs have recently been introduced onto UK motorways. Cameras read and record vehicle registrations throughout the network and it is then a simple matter of averaging the times of all the vehicles who were recorded at the origin and destination. They are therefore real time indicators of the journey time ahead.
(Handy little system if the Government ever wanted to automate toll charges for use of the motorway network!)
posted by hmca at 4:40 AM on May 2, 2008
(Handy little system if the Government ever wanted to automate toll charges for use of the motorway network!)
posted by hmca at 4:40 AM on May 2, 2008
This thread is closed to new comments.
posted by GuyZero at 5:50 PM on April 29, 2008