wires on road

tabatha

can anyone tell me what the wires lying on the belgard road are? i have noticed that there are a couple of them in different locations. there are two black ones running from one side of the road to the other and then conected to a box of some sort. there is one of the just beside jacobs i think and another just after the luas line. any ideas?

copacetic

they are for tracking traffic volume and speed.

tabatha

oic, thanks for that!

cast_iron

How accurate are they?

I mean, can they differentiate between a 5 axle truck as opposed to a single car.
Or is the percentage of trucks to cars recorded, and then used to decipher the difference?

Wishbone Ash

http://www.metrocount.com/products/mc5600/mc5600series.html

seamus

cast_iron wrote:

How accurate are they?

I mean, can they differentiate between a 5 axle truck as opposed to a single car.
Or is the percentage of trucks to cars recorded, and then used to decipher the difference?

I'm not sure, but it definitely has to be possible to work out what kind of vehicle is rolling over the cables. Think about the weight of the vehicle, the distance between the tyres, that kind of thing. Thinking about it now from a programming POV, it should definitely be possible to write software that would identify types of vehicles well into 99% accuracy based on the above.

cast_iron

seamus wrote:

I'm not sure, but it definitely has to be possible to work out what kind of vehicle is rolling over the cables. Think about the weight of the vehicle, the distance between the tyres, that kind of thing. Thinking about it now from a programming POV, it should definitely be possible to write software that would identify types of vehicles well into 99% accuracy based on the above.

I don't see how with this system. As I suspected, all it does is record an axle via a burst of air. It doesn't appear to differentiate between different weights or width of axles. It appears it needs a traffic monitoring system that records the truck to car to bike ratio; the avg num of axles per truck; rush hour times, etc.

Only then would the software be useful. Or am I picking it up wrong?

copacetic

cast_iron wrote:

I don't see how with this system. As I suspected, all it does is record an axle via a burst of air. It doesn't appear to differentiate between different weights or width of axles. It appears it needs a traffic monitoring system that records the truck to car to bike ratio; the avg num of axles per truck; rush hour times, etc.

Only then would the software be useful. Or am I picking it up wrong?

i would have thought all you need is the sensor, in software you could work out the type of vehicle just off the sensor data. however by the sounds of it this is built into the unit they have at the side of the road based on the weight and distance between axles.

Adaptive auto-ranging axle sensors detect bicycles through to heavy vehicles, eliminating sensitivity adjustments, end valves, bleed holes and any other periodic user calibration.

cast_iron

daveym wrote:

i would have thought all you need is the sensor, in software you could work out the type of vehicle just off the sensor data. however by the sounds of it this is built into the unit they have at the side of the road based on the weight and distance between axles.

Adaptive auto-ranging axle sensors detect bicycles through to heavy vehicles, eliminating sensitivity adjustments, end valves, bleed holes and any other periodic user calibration.

But the sensor here appear to be just a couple of pneumatic tubes that pump a burst of air back to the box at every passing wheel. The box just records the number of air packets (and presumably a time stamp).

Would the traffic data not still be needed to make sense of this? I can't see how the above data alone can tell the difference between various vehicle types.

The accompanying broshures on the quoted website actually amount to saying nothing except air bursts are recorded.

Tenshot

The NRA's website has a lot of statistical data gathered from road sensors here:

http://www.nra.ie/Transportation/TrafficDataCollection/TrafficCounterData/

(I think they move the sensors around, since the data listed ranges from 2002 to 2006 depending on which road you look at.)

copacetic

cast_iron wrote:

But the sensor here appear to be just a couple of pneumatic tubes that pump a burst of air back to the box at every passing wheel. The box just records the number of air packets (and presumably a time stamp).

Would the traffic data not still be needed to make sense of this? I can't see how the above data alone can tell the difference between various vehicle types.

The accompanying broshures on the quoted website actually amount to saying nothing except air bursts are recorded.

wouldn't the speed of the burst and the distance between them indicate the weight of the vechile and the distance between axles which would make them easy to autolog? they mention autoranging sensors so I assue this is what they do. They are used everywhere to record traffic stats including vechile type...

Winters

They are used for measuring traffic link flows and can differentiate between HGV and cars by way of axel count. Primarily used by the NRA, local authorities or private contractors like Abacus, Count On Us or Sky High.

Alun

There are a number of PDF's available on that website that give a hint as to how it works. Firstly, it just records axle hits, full stop. No information about vehicle weight that I can see, but it potentially (I can't see anything regarding this explicitly) it could record the length of the pulse. That combined with the speed, obtained independently due to there being two sensors on the road, could give some idea as to the size of the tyre footprint, and therefore extrapolate info regarding the size of the vehicle I suppose. Otherwise, for the unit referred to, no processing at all is done at the roadside, it just stores information about each axle 'hit', and all the processing is done afterwards on a PC. There's mention in the blurb about the software of 'classification' which I guess takes common patterns of axle bursts and classifies them into categories of vehicles. It might not be 100% accurate, but I assume they've done enough research into this to be pretty certain that their results will be accurate to a certain level. I suppose you could email them and ask, but it could well be proprietary information.

cast_iron

daveym wrote:

wouldn't the speed of the burst and the distance between them indicate the weight of the vechile and the distance between axles which would make them easy to autolog? they mention autoranging sensors so I assue this is what they do. They are used everywhere to record traffic stats including vechile type...

I can't see any way it can figure out the vehicle weight. A heavier vehicle will not send the air burst any quicker. As for the distance between axles, how does it know which is the last axle of a truck as opposed to the 1st axle of a following car. If traffic is moving slowly, this would be impossible to figure out.

Alun wrote:

but it potentially (I can't see anything regarding this explicitly) it could record the length of the pulse. That combined with the speed, obtained independently due to there being two sensors on the road, could give some idea as to the size of the tyre footprint, and therefore extrapolate info regarding the size of the vehicle I suppose.

I don't think there is a set distance between the sensors. It would have to be very accurate if it was going to figure out the tyre foot-print, and more importantly, like you say, they don't mention anything about this in their brochure (and they would if it was possible).

Alun

cast_iron wrote:

I can't see any way it can figure out the vehicle weight. A heavier vehicle will not send the air burst any quicker. As for the distance between axles, how does it know which is the last axle of a truck as opposed to the 1st axle of a following car. If traffic is moving slowly, this would be impossible to figure out.

What I meant was that a vehicle with a larger tyre footprint, like a truck, might cause a longer burst of air than a car or a bike. That combined with speed measurement (see below) might allow you to distinguish between different classes of vehicle. I'm not sure though whether the system actually records the length of the pulse though, or just the time that it occurs.

Anyway, I dare say that after several decades of experience in the field, they've got a few clever tricks up their sleeves regarding making sense of this kind of data which we can only guess at here.

I don't think there is a set distance between the sensors. It would have to be very accurate if it was going to figure out the tyre foot-print, and more importantly, like you say, they don't mention anything about this in their brochure (and they would if it was possible).

There's anther PDF on the site discussing the accuracy of speed measurements, and the effect of various factors including how accurately the distance between the two sensors has been measured and entered into the software. The distance isn't set, as such, but it can be measured after installation and entered as a parameter.