Those black traffic cables

August 10, 2007 7:46 AM

Those black cables for measuring traffic: what are they called?

You know, the ones that stretch across the pavement, they're tacked down on one side and run into a little metal "toolbox" on the other. What all exactly are they capable of measuring? Are they standalone or networked? Just curious--they're popping up like weeds in my neighborhood since the bridge collapse.

posted by gimonca to technology (12 comments total)

AFAIK, a single cable measures traffic volume and double cables measure speed (and volume).
posted by caddis at 7:53 AM on August 10, 2007

They measure traffic by counting the number of cars that drive over them. You know those rubber hoses at the gas station that ring the bell when you drive over them? Same principle.

By putting 2 cables a few yards apart, you can distinguish between traffic in each direction: If cable A is activated before cable B, then you're driving north. If B is activated before A, then you're driving south.
posted by cribcage at 7:54 AM on August 10, 2007

They're called "pneumatic road tubes".
posted by smackfu at 8:14 AM on August 10, 2007 [1 favorite]

The pneumatic systems are mostly used for raw traffic/axle counts.

There are a variety of other types of systems which are used for more extensive classification studies, and they are usually (but not always) incorporated directly into the road.

Specifically, piezo electric elements are embedded in pavement in pairs to sense vehicle speeds, as well as axle counts. Inductive loops are often used in conjunction with them (in piezo/loop/piezo configuration) to determine vehicle classification. (It's not enough to know the axle count... you need to know how the axles are grouped to determine what class of vehicle went over the sensors.)

Bending plates and quartz piezo sensors are used to weigh and classify vehicles while in motion. (This is a tricky problem, and one which I have been working on for the last few years, oddly enough.)

There are lots of these types of systems installed in interstate highways, but if you don't know what to look for, you'd not know you were being weight, counted, and classified.

Around your bridge failure, the pneumatic systems are probably collecting traffic preference patterns in support of the rebuild effort for the bridge.
posted by FauxScot at 8:40 AM on August 10, 2007

They are stand-alone. Generally they are deployed for short periods of time to do isolated traffic studies (typically volume, especially if there is only one tube).

Most transportation agencies do have networked traffic detection however. This can come in the form of installed loops (look for rectangular or circluar sawcuts in pavement) along highways (if you see them near traffic lights then they serve the purpose actuating signal-timing) or video detection. These systems can measure speed, traffic volume, and traffic density. Video detection is sometimes capable of classifying vehicles (especially commercial vehicles)
posted by nameless.k at 8:46 AM on August 10, 2007

In the US they're just used for traffic counting. In some other countries (South Africa, for example) they often also have a camera attached to them are are used for speeding tickets by measuring the time it takes to cross both tubes.
posted by chundo at 8:56 AM on August 10, 2007

Sorry to piggyback, but I've got a followup question that I think will be relevant/interesting for the OP and posterity:

Since we've apparently got a couple of traffic measurement engineers in the room, do any of you know if and how raw, realtime traffic data is made available to the public? Obviously the pneumatic sensors aren't pollable, but the loops are networked and are apparently made accessible by some means, since that's gotta be where Sigalert.com, Google Maps, MS Live Search etc. get their info. Have these companies made some backroom deal with the DOT (or whoever maintains the sensor network,) or is there an application process, or is this data just freely published somewhere?

Being able to scrape that info for use in homebrew projects sounds like loads of fun.
posted by contraption at 9:22 AM on August 10, 2007

See also prior AskMeFi post, What are the black strip on the road?
posted by sdinan at 10:51 AM on August 10, 2007

Especially using the example that chundo cited, how do you measure a specific vehicle's speed given the difference in wheelbases among various makes and models? I understand how you could approximate it in the aggregate, given an "average" wheelbase measurement, but trying to measure the speed of a single car based on these things seems a bit problematic.
posted by Doofus Magoo at 11:30 AM on August 10, 2007

If you only had one tube, you would need to know the wheelbase to measure speed. If you have two tubes offset a bit, you don't need to know it, since you can measure the delay between the front wheels crossing each tube.
posted by smackfu at 11:51 AM on August 10, 2007

Contraption:
The information is technically in the public domain, however I'm not aware of any agencies that are currently advertising access to the raw data as it is generated. In general, companies that publish this data on the web (outside of the generating agency's website) are not paying for it and have struck up some kind of deal. Same with the 'traffic cam' footage you often see on local news (video streams maintained by the DOT but technically owned by the public).
posted by nameless.k at 12:21 PM on August 10, 2007

Here's a vehicle record showing what some of the data looks like from a weigh-in-motion site:

-------------------------------------------------------------------------------
37163 68 mph Lane: 1 2007.06.26 22:43:08 Temp:68 Viol:00
Ve.-Code: 9 Total 1 2 3 4 5
Weight right (kips) 35.3 7.5 6.4 6.2 7.0 8.3
Weight left (kips) 34.4 5.2 6.1 6.6 7.5 9.0
Weight (kips) 69.7 12.7 12.5 12.8 14.5 17.3
Spacings (feet) 72.3 19.1 4.3 34.6 5.2
------------------------------------------------------------------------------

This vehicle has 5 axles and is a "Class 9" vehicle. Weights are composite, per-axle, per wheel. Vehicle speed is computed from the time delay from sensor 1 to sensor 2, separated by 12.0 feet, nominally. Length and interaxle spacing can be deduced from the speed, once known. Accel/decel can be deduced from changes in speed from sensor 1 to sensor 2, once length is known.

Three sensors were involved in this measurement (piezo-loop-piezo) and a temperature reading.

The output of the loop sensors is derived from changes in oscillation frequency from a base frequency that changes when a mass of metal is in the loop. Piezo signal amplitudes are related to the amount of weight applied, but it's non-trivial to make this measurement. A number of environmental factors influence the amplitudes.

Who'd a thunk all this crap is going on the interstates? Tech is everywhere.
posted by FauxScot at 1:07 PM on August 10, 2007