Slashdot Mirror

Wouldn't it be a lot easier to mount sensors on stop lights or buildings? You'd get a more consistent measurement. That is probably not what this is about. If you can't related it to self-driving cars, it can't be the next big thing. Innovation in Silicon Valley is petering out.

If we compare the number of the 250 Google street view cars to the 311,000 traffic signals in the United States we can easily see that your suggestion would require a lot more sensors.
Add to that that your suggestion requires that someone goes to the traffic signals to mount the sensors while the cars can be modified when they are in for service.
The project in the article could probably be done for $250,000 or so depending on the cost of the sensor and how they are planning to mount it.
I don't see how your suggestion could possibly be done for less than $100,000,000.

specifies that the duration of the yellow change interval should be between 3 and 6 seconds. And people have won court cases over red light tickets over the yellow time being too short.

http://www.ite.org/decade/pubs/IR-117-E.pdf

http://www.ite.org/safety/issuebriefs/Traffic%20Signals%20Issue%20Brief.pdf

http://www.ite.org/annualmeeting/compendium10/pdf/AB10H2601.pdf

specifies that the duration of the yellow change interval should be between 3 and 6 seconds. And people have won court cases over red light tickets over the yellow time being too short.

http://www.ite.org/decade/pubs/IR-117-E.pdf

http://www.ite.org/safety/issuebriefs/Traffic%20Signals%20Issue%20Brief.pdf

http://www.ite.org/annualmeeting/compendium10/pdf/AB10H2601.pdf

specifies that the duration of the yellow change interval should be between 3 and 6 seconds. And people have won court cases over red light tickets over the yellow time being too short.

http://www.ite.org/decade/pubs/IR-117-E.pdf

http://www.ite.org/safety/issuebriefs/Traffic%20Signals%20Issue%20Brief.pdf

http://www.ite.org/annualmeeting/compendium10/pdf/AB10H2601.pdf

I recently went to an ITE luncheon at which the LAPD presented a (surprisingly thorough) study of various red light camera countermeasures on the market. If they were telling the truth in their presentation, then not one of those products works worth a hoot. They had better results with clear acrylic paint than with covers or the specialty paint.

It's probably worth noting that the cameras in question were fairly new.

There are about 100 vehicle related fatalities in the United States everyday. 40,000+ a year.

http://www.ite.org/crashes/index.htm

(just an interesting statistic, not really a counter to anything you said)

I work in the traffic engineering industry. There are two things you have to realize about speeding: first, many speed limits are set artificially low and second, speed itself isn't dangerous--it's the difference in speed that causes accidents.

The accepted method for setting speed limits is to collect speed data on all vehicles on the road for 24 hours on a typical day. This is usually done using those two rubber tubes you may see placed across the road at times. The speed limit is then supposed to be set at the 85th percentile speed of the traffic. In reality, political considerations often force the speed limit to be something ridiculous like 55 mph. If 85 percent of the traffic is doing 70 mph, that speed limit is unrealistic and will be ignored.

The other part of this is the common misconception that slowing the traffic down will make everything safer. While it is true that accidents that happen at higher speeds are more severe, accidents are caused by difference is speeds. Studies have shown that the number of accidents sharply increases when the differences in speeds between vehicles exceed 20 mph. This means that if you are the guy driving 55 mph when everyone else is doing 80, you are the one driving recklessly. Because of this fact, artificially low speed limits can actually cause accidents.

This standard from the Institute of Transportation Engineers sums everything up nicely: http://www.ite.org/standards/speed_zoning.pdf

You're legally right about the speed limit being an upper limit, but in most jurisdictions speed limits are set artificially low for a variety of reasons: complaining by people who don't know better, revenue from traffic tickets, etc. There is actual science and engineering knowledge of how to do this (http://www.ite.org/standards/speed_zoning.pdf), but they of course science, knowledge, and curiosity in general get roundly ignored by ignorant and pandering politicians, soccer moms, and the rest of the Think of the Children types.

Interestingly, people who drive faster or slower than the prevailing speed (regardless of speed limits) are more likely to be involved in accidents than those who go with the flow of traffic. Of course, highway robbers (police and politicians) don't focus on people who drive too slowly. So maybe I too should take that anger management class you mentioned, but I'm getting sick of people pretending reality is something that it's not. Set speed limits according to sound scientific principles and things will be a lot better.

I don't think the traffic light them selfs are running on x86 hardware they more likely running on a Embedded system and yes they are networked / linked to other lights, railroad crossings, and some times back to some control center.
The camera system may be running x86 hardware.

and you want source code and/or the "version info"/ bugfix list, along with the version for the lights and the camera system as well the light timing settings. Also some times LED lights have a some lag before they are fully on this is more likely on light that used to be the non led type that have led installed in to them. Also some cameras go off on amber / Yellow and some the amber time is too short for the speed limit.

http://www.thenewspaper.com/news/16/1621.asp
http://www.ite.org/standards/atc/
http://www.trafficsignalmuseum.com/

You know, I've never worked in NYC, but I've visited, and one thing I recall is the following: the yellows go to red, and then there's another couple seconds of pause, and then the reds on the other side flip to green. Also, many members of ITE (Institute of Transportation Engineers, the traffic engineer equivalent to IEEE) have been bitching about NYC's idiotic 3 second yellow standard for a long time now. Let's also note that NYC's drive to a 3 second yellow was spearheaded by an insurance industry group that also pushes red light cameras (and that the far too short 3 second yellow has been proven to cause higher numbers of entry-on-red, which results in higher ticket revenue). Something that causes higher numbers of crashes and tickets, being pushed by an insurance industry lobbying group? Gee, you don't think they might want to be able to jack up their rates for drivers, do you?

My aggressiveness comes from the fact that I spent most of my high school summer jobs working as a surveyor's assistant, meaning that I dealt with more than a few road engineers. I'm not dumb, and I picked up quite a bit. Having a civil engineer for a father probably helped too. Oh, and knowing of the existence of the Highway Safety Act (1966) and the Manual of Uniform Traffic Control Devices. Which used to lay out exactly the principles I talk about, until it was modified in 2000 to remove the "based on formal site engineering studies" for politically motivated reasons (specifically, the lobbying by red light camera and insurance industry people, and also the desire by most cities for increased revenue from red light cameras). However, even the watered down version states the reason for a yellow light to be (paraphrasing) "to allow motorists adequate time to stop prior to the intersection and to clear the intersection." ITE/ASCE's most recent publication setting forth the method of determination would be Determination of Left-Turn Yellow Change and Red Clearance Interval, published in JTE. Specifically it deals with those issues for left-turn intersections. Unfortunately, the paper is not free, unless you have access to JTE through work or school or membership. Another non-free reference is ITE's #JDB04A20

Be less aggressive when you try to talk back to someone who knows what they're talking about, okay? NYC is hardly the typical example for road safety, by the way - you freaks don't even obey your lane lines, much less your traffic signals.

We already *have* freeways. Traffic engineers should work on slowing cars on surface streets.

Don't worry... they are! Search for "Traffic Calming" and you'll find things like:

http://www.trafficcalming.org/

http://www.ite.org/traffic/

http://www.fhwa.dot.gov/environment/tcalm/

Well, for one thing, drunk drivers aren't going to be caught by RFID, unless they're speeding which they don't need to do to be dangerous.

You might want to look into the literature for traffic calming (also see trafficcalming.org). There's plenty of evidence that simple measures can dramatically reduce the number of casualties.