It's almost like if you get together a talented team of people, give them a clear brief and scope and then let them use whatever tools they need you end up with a good product at the end...
Thanks for posting this. My city is getting ready to start putting in red light cameras, with the supposed goal of improving traffic safety. This should come in handy when they optimize for revenue.
I was thinking about this the other day. Many articles claim this happens, but few include real world before/after measurements.
You should take timings + videos of pre- and post-camera activity, for future reference! It might make a nice compliment to yet another 'they shortened yellow light timing' article.
I wonder how many man-hours it took to develop this document. It is thorough and easy enough for a layman to understand. If we had such specifications when building software systems, life of programmers would be much easier.
The word 'bicycle' is mentioned only two times. 'bicyclists' once, 'bike' none. Why are we officially overlooking a fantastic and common form of transport?
The primary author of this document is Peter Koonce, who is a pretty serious bicycle guy. You might be interested in his blog: http://koonceportland.blogspot.com/
One of the local tv stations here did a great story on traffic light phasing.
They first had viewers email them their worst bottlenecks and where they felt traffic lights hampered their commutes. After several weeks of collecting angry emails, they narrowed it down to a handful of intersections.
Then they interviewed a local DOT person, and then brought them out to the worst intersection the viewers identified. The reporter said they spent about 30 minutes working on the lights and signaling. The reporter did say that traffic was flowing better and there were less cars backed up at the light after the DOT person corrected some of the signaling issues.
Of course they ended the story by lamenting about how many traffic lights there are in the city and clearly the DOT have an uphill battle in keeping every intersection flowing well, but said they spend a lot of time at least trying to monitor the bad intersections.
GREAT FIND! Very interesting read. I liked the parts about how lack of concern for pedestrians in the system is a legacy issue. Also interesting is the segment on stop bars for high speed stops. I never considered that the light stays yellow extra long to let cars clear! I've always thought "Good choice Phil, you got through on the yellow all the way." It appears that some of those lights stay yellow until the car is through! (TSTM Ch 4: 4.7.4)
I've fantasized about making a more responsive, consistent traffic light for years. Every time it's the middle of the night and I'm sitting at a red light for no reason --- or when I figure out that the sensors are broken and I need to get out of the car and press the cross walk to have the light change, or when I'm on my bike and have no chance at all to trigger the sensors.
All these things I want fixed. I want stop lights to stop sucking.
I want an unprotected left to have the signal on the opposing side turn red sooner if there's a lot of people in the queue ... so that more than two cars can turn per green.
There's countless improvements that could be done but aren't that are purely in software and would use the same signal installations. We're not moving mountains here.
But clearly nobody has done this. Stoplights suck exactly the same ways, in exactly the same amounts, as they did 25 years ago.
And then, instead of making them suck less, we slap cameras on them to make sure that people put up with the suckage or pay a stiff penalty. "This technology has annoying bugs that have never been fixed. But luckily, we've just spent a lot of money on making sure you don't work around them."
Who does that? This is absurd. no new features - just fix the damn bugs.
The problem with making the opposing red end sooner ("oncoming traffic has longer green," as the signs say) is that the opposing permissive left turns that haven't cleared by the end of the yellow are then caught in the intersection with their signal red and yours still green. The problem is called the "yellow trap" (http://en.wikipedia.org/wiki/Yellow_trap), and some states, including California, outlaw the practice because it is too dangerous.
I'm completely with you about wanting everything demand-actuated to be very responsive, though. Traffic signals are all engineering with no user interface design at all.
A recent trip to the UK was enlightening for me. Their roadways have a fraction of the traffic lights and stop signs that we have on the east coast. We need to remove stoplights from lightly trafficked intersections. For heavy use intersections, replace with rotaries. The glut of suburban stop signs need to be replaced with yield signs.
19 comments
[ 428 ms ] story [ 1070 ms ] threadThe people behind it are incredibly talented, they have a blog here https://gds.blog.gov.uk/
It's almost like if you get together a talented team of people, give them a clear brief and scope and then let them use whatever tools they need you end up with a good product at the end...
Also good https://digitaltransformation.blog.gov.uk/
Also interesting https://www.gov.uk/service-manual/agile/index.html which is their manual for providing digital services.
You should take timings + videos of pre- and post-camera activity, for future reference! It might make a nice compliment to yet another 'they shortened yellow light timing' article.
They first had viewers email them their worst bottlenecks and where they felt traffic lights hampered their commutes. After several weeks of collecting angry emails, they narrowed it down to a handful of intersections.
Then they interviewed a local DOT person, and then brought them out to the worst intersection the viewers identified. The reporter said they spent about 30 minutes working on the lights and signaling. The reporter did say that traffic was flowing better and there were less cars backed up at the light after the DOT person corrected some of the signaling issues.
Of course they ended the story by lamenting about how many traffic lights there are in the city and clearly the DOT have an uphill battle in keeping every intersection flowing well, but said they spend a lot of time at least trying to monitor the bad intersections.
Several state DOTs have endorsed the NACTO USDG.
All these things I want fixed. I want stop lights to stop sucking.
I want an unprotected left to have the signal on the opposing side turn red sooner if there's a lot of people in the queue ... so that more than two cars can turn per green.
There's countless improvements that could be done but aren't that are purely in software and would use the same signal installations. We're not moving mountains here.
But clearly nobody has done this. Stoplights suck exactly the same ways, in exactly the same amounts, as they did 25 years ago.
And then, instead of making them suck less, we slap cameras on them to make sure that people put up with the suckage or pay a stiff penalty. "This technology has annoying bugs that have never been fixed. But luckily, we've just spent a lot of money on making sure you don't work around them."
Who does that? This is absurd. no new features - just fix the damn bugs.
I'm completely with you about wanting everything demand-actuated to be very responsive, though. Traffic signals are all engineering with no user interface design at all.
http://www.cs.utexas.edu/~aim/
I did get a chuckle out of the header: "21st century operations using 21st century technologies". The lady doth protest too much, methinks.