Too many traffic lights?

[Duke87]

Read this article.

Now, for a little background. Veera and Mani? They're coworkers of mine. This Mitch Mailman guy? He's a nutcase, and we've been dealing with him complaining for weeks now. He even came into the office with his son, to use as a sort of dramatic example piece ("my son here is afraid to cross the street at that intersection!"). And when my boss saw straight through the tactic and calmly pointed out that we had data and traffic studies while he had no numbers to back his claims up, he started making a scene, yelling about how obstinate the city was being. We almost had to call security to remove him.

But on the matter at hand... are there cases where, perhaps, a signalized intersection would be better served by a stop sign?

I'd say, generally, no. Not a multi-way stop sign, anyway. Yes, it makes everyone stop, and yes that does in theory increase safety since no one is ever going through the intersection at free flow speed (unless they run the stop sign...), the problem is that when every car has to stop, it really limits the rate at which they can proceed through, and if you do it on too busy a road, it can really create gridlock. Hence the signal. Now, if we're talking about letting the main street free flow and just giving the side street a stop sign... then maybe in some cases it would be better. A traffic circle could also work, but that doesn't work when the street gets too crowded, nor is it optimal to force mainline traffic through them (textbook situation for a traffic circle is a three way junction where all three movement pairs have significant volume).
Ultimately, if lights are creating driver frustration, the problem is not that there are too many, it's that they are poorly timed. You shouldn't be catching three consecutive red lights. There can be a light at every driveway and it's not a problem so long as you can go through long strings of them in a row green.

Your thoughts?


(on another note, there's a slight error in the article: the 50 and 115 second values given are the length of a full cycle of the light, not the length of a red phase)

[Bryant5493]

Low traffic volumes and minor streets work being an all-way stop intersections, if people pay attention. Major streets with high traffic volumes don't work as all-way stops. It just becomes gridlock, like you said. Also, too high of a concentration of red lights, out of sync, is just as bad... maybe worse.

http://www.youtube.com/watch?v=HJqu5rzGIpc

This is Tara Blvd. (U.S. 19/41/S.R. 3) in Clayton County, Ga. It's traffic light hell.


Be well,

Bryant

[74/171FAN]

That's the case along VA 36 in Hopewell between VA 144 and I-295.  Barely ever will you get through all three signals without stopping at one or maybe two of them

[froggie]

Duke:  since you mention the subject, MnDOT recently removed a traffic signal from MN 19 in Marshall.  The signal warrants were no longer being met, so it was replaced with side-street stop control.

[Bickendan]

The article mentioned that a walk signal being pressed will add 30 seconds to the cycle. Here's a solution that may work: Remove the walk signal buttons. When the signal is green, the walk sign is on. Within 10 seconds, have the stop crossing light flash with a countown timer. This will tell both the pedestrian how long they have to finish crossing and it will tell motorists how long the signal will remain green until it switches to yellow (at zero).

This will also have the bonus of preventing pranksters of hitting the walk signal when they have no intention of crossing.

Secondly, I disagree with traffic circles being ineffective. I go through the circle at NE 39th Ave and Glisan St in Portland on a regular basis, and traffic flows quite well throughout the day, despite the roundabout nature and the stop signs (not signals) controlling traffic entering the circle. 39th Ave is Portland's 'midtown' north-south artery, connecting the Hollywood District and I-84 to the north to the Hawthorne District to the south. http://maps.google.com/?ie=UTF8&ll=45.526474,-122.623032&spn=0.001049,0.002473&t=k&z=19
I think the biggest problem that traffic circles have are their infrequency of use in the US and their ofttimes inappropriate configuration (I believe more than two lanes per direction cuts their effectiveness by promoting weaving).

[Duke87]

The article mentioned that a walk signal being pressed will add 30 seconds to the cycle. Here's a solution that may work: Remove the walk signal buttons. When the signal is green, the walk sign is on. Within 10 seconds, have the stop crossing light flash with a countown timer. This will tell both the pedestrian how long they have to finish crossing and it will tell motorists how long the signal will remain green until it switches to yellow (at zero).

That's how it works downtown... but we still have the buttons anyway. Elsewhere, the walk signal does not show up unless somebody presses the button, and when it does, it's a protected pedestrian phase (red for all traffic). People generally don't bother pushing the button and just cross cross the street as soon as there's an opening. So yeah, they are kind of pointless.

That said, we're not at liberty to change the way a signal is phased without CONNDOT's approval (even if it isn't on a state highway), and I do believe having pedestrian buttons is mandated even if the walk signal comes on automatically. They're not going to let us get rid of them.

[SSOWorld]

The article mentioned that a walk signal being pressed will add 30 seconds to the cycle. Here's a solution that may work: Remove the walk signal buttons. When the signal is green, the walk sign is on. Within 10 seconds, have the stop crossing light flash with a countown timer. This will tell both the pedestrian how long they have to finish crossing and it will tell motorists how long the signal will remain green until it switches to yellow (at zero).

That's how it works downtown... but we still have the buttons anyway. Elsewhere, the walk signal does not show up unless somebody presses the button, and when it does, it's a protected pedestrian phase (red for all traffic). People generally don't bother pushing the button and just cross cross the street as soon as there's an opening. So yeah, they are kind of pointless.

That said, we're not at liberty to change the way a signal is phased without CONNDOT's approval (even if it isn't on a state highway), and I do believe having pedestrian buttons is mandated even if the walk signal comes on automatically. They're not going to let us get rid of them.

now that's just rediculous (the DOT having such control over a local road's traffic light)

I came across a traffic light that has a 4-way red while peds cross system.  Of course, no right turns were allowed on red either - two of the directions had stop lines 25 feet away from the actual intersection (protecting a gas station's driveway entrances 

[Bryant5493]

^^ I saw something similar a few years back when I was in Savannah (the Barnes Dance). It think that it was Bay Street, and another street, just south of River Street.


Be well,

Bryant

[Revive 755]

I'd say, generally, no. Not a multi-way stop sign, anyway. Yes, it makes everyone stop, and yes that does in theory increase safety since no one is ever going through the intersection at free flow speed (unless they run the stop sign...)

The St. Louis, MO area is a good example of what can happen when there are too many unwarranted all-way stops.  Safety gets worse because drivers will start to only slow down slightly, if at all, and then roll through the intersection.

I came across a traffic light that has a 4-way red while peds cross system.  Of course, no right turns were allowed on red either - two of the directions had stop lines 25 feet away from the actual intersection (protecting a gas station's driveway entrances

There are two intersections in Iowa City, IA at the Old Capital - both on Iowa Avenue - that have a ped only phase.  But since it takes awhile for this phase to come up, most peds cross when the mostly parallel direction of traffic gets a green light.  There is (was?) an intersection on the Ball State campus in Muncie, IN, that had an all-ped phase that also had ped signal heads for diagonal movements.

I think Harrisburg, PA, had some intersections where the ped phase came up a second or two before parallel traffic got a green.

Duke:  since you mention the subject, MnDOT recently removed a traffic signal from MN 19 in Marshall.  The signal warrants were no longer being met, so it was replaced with side-street stop control.

Back in July of 2007, there were several intersections in Topeka, KS, with stoplights signed for removal.  The north end of MO 267 near the St. Louis city limit also seems to have had some stoplights removed after the '93 flood - though MO 267 was sometime afterwards restriped to two lanes from four.

If we're eventually going to nominate roads with too many traffic lights, I'll nominate MO 141 between Valley Park and US 40 (see one of my many rants about it) and MO 30 from US 50-61-67 to somewhere around the Rte MM/Rte W intersection.  MO 30 is a case where more access consolidation, right-in/right-out only movements, and/or Michigan left type access should have been used instead of sometimes having a separate light to access a shopping center.

[vdeane]

I agree with the removal of the walk buttons.  The signal should always be on walk when the light is green unless there is a protected right turn phase.  In fact, why not remove the pedestrian signals altogether?  Most of the time the directions for cars and pedestrians are the same (except in cases where the signal is never on walk unless the button is pushed; these cases should be eliminated), so why spend money on extra stuff?  Drivers are supposed to watch for pedestrians in the intersection anyways.  This would also have the side effect that towns would be forced to have the yellow phase be of a decent length and would be unable to shorten it to increase tickets.

[Bickendan]

Leave the walk signals. The flashing don't walk phase is an indicator for the drivers that the green phase is about to end, particularly when their is a countdown timer.

[roadfro]

More importantly, pedestrian signal heads are used to give pedestrians a clear indication as to how much time is available for crossing.  Staring at only a circular green signal, a person walking has no indication of whether that green will remain for another 5 seconds or 55 seconds. The flashing don't walk tells a pedestrian that they don't have enough time to start crossing the street.  (Flashing don't walk also generally has the unintended side effect of being an indicator of remaining green time to motorists, although this is not always accurate.)

The MUTCD already mandates a 3-6 second yellow time, and there are guidelines/formulas (not universally followed) for setting this.  Pedestrian clearance time (flash don't walk) is not a controlling factor for the yellow interval.  Where pedestrian push buttons/signal heads are not present, the vehicular phase (either green or green+yellow) must accommodate the pedestrian crossing speed of 4 ft/sec.

[vdeane]

In Rochester NY they're useless as many never go to walk unless the button is pushed, making it so that it's usually faster to jaywalk .  Anyways, drivers are supposed to look anyways, both for pedestrians and to make sure that there aren't any idiots running the light.

[roadfro]

That is the purpose of the button...to activate the walk phase.  Note that a button press in the middle of the adjacent green phase usually won't activate the walk signal, as there may not be enough time remaining during the actual green time (not counting phase extensions) to allow for the ped phase.  Modern signal controller technology allows for pedestrian heads to activate during every green ("rest in walk" function)--I believe this is more commonly seen in coordinated signal systems not is universally applied.

Regardless of pedestrian signal mode, it's always faster to jaywalk.

[vdeane]

Wouldn't it make sense to have it say walk when green though?  Or maybe I'm just reasoning like that because in my area all signals are timed except at night (doesn't seem to matter when in NY you go with this either); some are timed even in the middle of the night.

[roadfro]

Wouldn't it make sense to have it say walk when green though?  Or maybe I'm just reasoning like that because in my area all signals are timed except at night (doesn't seem to matter when in NY you go with this either); some are timed even in the middle of the night.

Actually, if you've got a signal operating in the fully-actuated mode (or even semi-actuated on side streets), having the walk come on during every green phase would result in the green lasting longer than necessary.

Example: A signal is semi-actuated and the walk phase is used during every green phase. Lets assume the the main street has five 12-ft lanes (two each way plus a turn lane) for a total width of 60 feet. Also assume a 4-sec yellow and 1-sec all-red time for the signal.  When a single car pulls up on the side street, the main street will turn red and the side street green phase comes on.  In our scenario, the side street walk signal also comes on automatically, despite the fact that there are no pedestrians crossing. If the walk signal automatically activates with this side street green phase, this will result in a side street phase that lasts for a minimum 24 seconds: 4 sec MUTCD minimum walk time + 15 sec flash don't walk time (60 ft @ 4 ft/sec MUTCD minimum crossing time) + 4 sec yellow time + 1 sec all-red time.  If the automatic walk is not used, the side street phase could be as small as 9 seconds (assuming a minimum green value of 4 seconds). Between these two situations, the automatic walk scenario interrupts main street traffic for 15 seconds longer--that interruption increases by 3 seconds for every lane added on the main street.

The only times where it makes sense for the walk signal to come on automatically is (a) when the signal operates in a fixed-time mode, because the conflicting green times won't change, or (b) on coordinated arterials where a major street will need to be green for a specific period of time to maintain coordination, and that green period can incorporate the pedestrian phase with no adverse affect on coordination.

[vdeane]

In NY almost all signals are fixed-time during the day, and it's very rare for any phase of a light to last less than 20 seconds (except for protected turns and a few special cases).

[roadfro]

In NY almost all signals are fixed-time during the day, and it's very rare for any phase of a light to last less than 20 seconds (except for protected turns and a few special cases).

I'm curious, are you referring to all of NY state or just NYC?  Using fixed-time is really inefficient signal practice, except in dense downtown grids.

[Hellfighter]

M-153 between Inkster and US-24/Telegraph Roads
Most streets in the City of Detroit
Old US-12/Old M-14 Plymouth Road between Burt Road and Greenfield Road. A light every 1/4 a mile!

[vdeane]

I'm curious, are you referring to all of NY state or just NYC?  Using fixed-time is really inefficient signal practice, except in dense downtown grids.

The state.  Here in NY, we're the last to get everything.  Also, NIMBYism is so prevalent here that most of our roads are above capacity - additional roads and widening are out of the question due to money and NIMBYs, and traffic lights are quite common except in rural areas.

[Duke87]

In NY almost all signals are fixed-time during the day, and it's very rare for any phase of a light to last less than 20 seconds (except for protected turns and a few special cases).

I'm curious, are you referring to all of NY state or just NYC?  Using fixed-time is really inefficient signal practice, except in dense downtown grids.

Well, I'll tell you this much: in New York City there are traffic lights which are still using mechanical control boxes (hey, at least they're EMP-proof!). I somehow doubt the signal system is too intelligent.
There must be thousands, probably tens of thousands, of Traffic signals around the five boroughs. Linking everything up and automating it would be a huge expensive ordeal. And the turnaround time on replacing things no doubt gets quite large at times.

[thenetwork]

Out in my neck of the woods,  the crosswalk signals, much like the the signals themselves, are all on a fixed time/synchronized citywide grid -- meaning that during the day, most intersections have a fixed time phase in each direction and the crosswalk lights are always activated on every green (usually the time of day when most pedestrians DON'T push the button).  Depending on the area, as evening sets in, signals go into a shorter phase cycle where the lights only change when there is traffic present.  Usually at this time as well, the crosswalk lights remain red (or DON'T WALK) unless someone actually pushes the crosswalk button.

Most crosswalks in our area now have countdown timers.  Those signals are also beneficial for traffic to tell roughly how quickly a light "could" change, since some lights don't necessarily cycle to red when the timer hits "00".  So in off-peak hours, when cross traffic is at an intersection, the light can change right away. 

Older signal intersections I've seen will activate a flashing DON'T WALK cycle of 10+ seconds before the light changes for the cross traffic.  By keeping the crosswalks at DON'T WALK in off peak hours, within those 10 seconds, the light could have already cycled for the cross traffic and perhaps returned to the original green phase for the mainline traffic, saving time and gas.

[SSOWorld]

I've come across a traffic light that - when no traffic is present, every direction has red lights 

[rawmustard]

I've come across a traffic light that - when no traffic is present, every direction has red lights 

Red rest is a setting on fully actuated controllers, but one I haven't seen in the field too often. Usually most jurisdictions will have a primary street in green rest, and some will merely rest in whichever phase was last activated

[agentsteel53]

I do not recall ever seeing red rest, but I may just have not noticed since I'm not good at observing the state of side-road signals.  Likely I just thought "hm, I wonder why I get red in the middle of the night even though I'm the through street".