Dr Zarkoff Wrote:
-------------------------------------------------------
> A correction, this:
>
> "which uses track circuits to determine track
> occupancy -- rails are detected as a track
> occupancies."
>
> should read:
>
> "which uses track circuits to determine track
> occupancy -- broken rails are detected as a track
> occupancies."

You are the man, dude, great information on these systems.

The freeway inductive loops use the same process to estimate the traffic speeds, that are processed through an algorytm to aggregate and average them across lanes, to populate all of those traffic maps on the internet. I had the job of calibating and verifying all of those Caltran's freeway loop detectors for those maps back in the early 2000's, and I love the people at Manpower who we brought in to help me. Worse than working the West Colton hump, hours of looking at satellite images of freeway segments and geolocating thousands of loops, until your eyes fall out, mind numbing....

They try to fine tune this a bit for public consumption but there is a plus or minus range with a short time lag until the next update of what you are seeing, that is tolerable for informational purposes when John Q Public is checking how bad there trip is going to be because of the truck accident. And the count/speed data is good for aggregate data analysis for demand for planning and such.

This bound of error is small enough, and due to micro changes, like more trucks in the midday than a at peaks, because and 18 wheeler operating at the same speed is going to seem like it is taking longer, than an average size car. So loops are "good enough" data inputs, for speed estimation on freeways for operational purposes. When the system managers see the loop detector permanently occupied for any significant length of time at the Caltrans/CHP TMC's, they can pull up the CCTV, or send the CHP to find out why the freeway is stopping and figure out what to do (send tow truck, send fire department, send major truck incident team, send coroner, etc.)....

Secret confession to many of you, the loop indications you are getting on the internet for arterials are mostly not worth a warm bucket of spit, due to things as fundamental as sample size to make the estimates or properly calibrate those smoothing algorithms... Some places now do it with visual detection, mounted on traffic signals, which are generally much more accurate.

There are programs that use visual imaging and count all types of vehicles, motor cycles cars, trucks, exact speeds, by lane, by individual vehicle, and literally "sample" the whole population of observed users. It allows us to look at operatiobal dynamcis under actual conditions as they change, even like how drivers change lanes in resoonse to varying merging volumes, varyis speeds/flow rates, like "breakdown" into stop and go conditions.

This data has allowed us to model actual freeway operation and see "what works and what doesn't" in our merge designs and interchange designs and reduce accidents and/or other conflict merges, which can desabilize traffic flow and reduce lane flow rates, this has been used to do things like set ramp meters more effectively, to optimize dynamic flow capacity.

I had no idea when I started my paid career with Stan and Bill and the Rogue River Number One, laying track(actually dragging around track jacks, pinch bars, heavy duty chains, and wire rope for dragging rails, behind Bill's pick up) and of course loading eucalyptus logs in the bunker of Number One when they were still in Alta Loma and Movie train was still a notion, that this was what I'd ever be doing in transportation 35 or so years later...