Traffic signal

[Scott5114]

Not too puzzling–since that is a T intersection, there's no situation in which any traffic would cross EB traffic. So there is no reason that a light for EB should ever turn red. Most places would at least put green-only signal heads there, but apparently in this case they decided not to bother.

[SignBridge]

You're right. I missed the right-turn only sign on the mast-arm the first time I looked at it.

[Big John]

That's very puzzling.

Thet would usually have the eternal green ahead arrows.

[jakeroot]

I think there needs to be a second left turn signal, per the MUTCD. One signal head for an approach hasn't been allowed in a long time.

[Big John]

I think there needs to be a second left turn signal, per the MUTCD. One signal head for an approach hasn't been allowed in a long time.

It is debatable whether that is a secondary signal.

[jakeroot]

I think there needs to be a second left turn signal, per the MUTCD. One signal head for an approach hasn't been allowed in a long time.

It is debatable whether that is a secondary signal.

It's the only signal at that intersection for that approach. Therefore it needs two signal heads. Most intersections have a through signal, which must always have two signals. But lacking a through signal, the left turn here would be the primary movement (I don't see the through movement as part of the intersection) and therefore needs two signals.

[Dirt Roads]

I think there needs to be a second left turn signal, per the MUTCD. One signal head for an approach hasn't been allowed in a long time.

It is debatable whether that is a secondary signal.

It's the only signal at that intersection for that approach. Therefore it needs two signal heads. Most intersections have a through signal, which must always have two signals. But lacking a through signal, the left turn here would be the primary movement (I don't see the through movement as part of the intersection) and therefore needs two signals.

There's a long-standing FRA regulation for railway signalling that applies here.  The rule is called "Light-out Protection" and, in a nutshell, you need to have a backup design if a burnt-out bulb would cause confusion that could upgrade the aspect of a signal.  To apply this to traffic signals, if a left turn signal has only one red arrow (or red ball) and it is burnt out, you MUST have a secondary signal (or a double red) if the dark signal might be missed and a green ball on the other heads could be confused as a protected left turn.  That doesn't happen very often, but it is certainly a problem for traffic signals complexes that once were programmed with green ball through and green arrow turn in the same phase (and I would make the safety case that this is also an issue if you are the lone signal with a separate left turn phase in a string of traffic signals with the other arrangement).

Keep in mind that the dark signal is the most restrictive aspect (both in railroading and traffic signals).  It is a great safety default if there aren't any other signal heads to get confused with.

[steviep24]

It seems that that particular entrance doesn't really need to be signalized at all given that traffic leaving the shopping complex can only turn right in the first place (no sign prohibiting left turns though). Unless they signalized it for pedestrian safety purposes.

A Whole Foods is under construction at this location which is probably why they did it. Still, the single FYA as the only signal doesn't make sense. As others have mentioned they could have the eternal green arrows there.

The other entrance to the complex on the other side of Starbucks is fully signalized and allows left and right turns.

https://www.google.com/maps/@43.112848,-77.5500978,3a,75y,120.81h,80.32t/data=!3m6!1e1!3m4!1sbw4FI0-6xf8wISk3MepqAQ!2e0!7i16384!8i8192?hl=en&authuser=0

[US 89]

It seems that that particular entrance doesn't really need to be signalized at all given that traffic leaving the shopping complex can only turn right in the first place (no sign prohibiting left turns though). Unless they signalized it for pedestrian safety purposes.

A Whole Foods is under construction at this location which is probably why they did it. Still, the single FYA as the only signal doesn't make sense. As others have mentioned they could have the eternal green arrows there.

I'd bet there's consistently very heavy oncoming traffic going northwest on Monroe. Having that FYA allows the possibility of a protected left turn for southeastbound traffic on Monroe into that complex.

That said, I don't think I've ever seen a perma-green setup like that where there wasn't some sort of curb or island or clear indicator that physically separated the unimpeded through lanes from turning traffic.

[jakeroot]

That said, I don't think I've ever seen a perma-green setup like that where there wasn't some sort of curb or island or clear indicator that physically separated the unimpeded through lanes from turning traffic.

Like this?

https://goo.gl/maps/VJmmHb2WYLUEGrG18

Or do you mean without at least a through signal?

[US 89]

Nah, I mean something like this or this or this. In all of those, there's a raised curb to the left of the through lanes to physically separate them and emphasize "no, these are not part of the intersection, you can blow through this light". The first example above used to not even have a green up arrow signal, but it didn't make a difference because the curbs made it clear it wasn't part of the signal-controlled intersection.

That Washington example would be #2 for me. Maybe the curb/island is just a Utah standard and I've just never paid all that much attention anywhere else.

[Amtrakprod]

New signal outside of peak hour:

In peak hour:



iPhone

[mrsman]

It seems that that particular entrance doesn't really need to be signalized at all given that traffic leaving the shopping complex can only turn right in the first place (no sign prohibiting left turns though). Unless they signalized it for pedestrian safety purposes.

A Whole Foods is under construction at this location which is probably why they did it. Still, the single FYA as the only signal doesn't make sense. As others have mentioned they could have the eternal green arrows there.

I'd bet there's consistently very heavy oncoming traffic going northwest on Monroe. Having that FYA allows the possibility of a protected left turn for southeastbound traffic on Monroe into that complex.

That said, I don't think I've ever seen a perma-green setup like that where there wasn't some sort of curb or island or clear indicator that physically separated the unimpeded through lanes from turning traffic.

While setting up a perma-green signal in such situations seems ideal to avoid driver confusion, I don't think those signals are strictly necessary.  Left turn at the signal in Rochester is controlled by an FYA signal, but straight through is uncontrolled.  Straight through traffic will simply go straight through and not be worried that they don't have a signal.  They simply proceed.

Jakeroot mentioned that MUTCD would likely require two left turn signal heads here as it is the primary movement.  I wonder if that would be true if the striaght through movement were signaled with perma-greens.  One left turn signal head and two straight signal heads, but the straight signal heads are all perma-greens.  So if only one left turn signal head is necessry if they have two perma-greens, why would more than one be necessary when the perma-greens are replaced with nothing, which is the functional equivalent?  I do agree that it is good practice to put in two left turn signal heads (IMO everywhere), but I don't think it is strictly required.

US 89 mentioned that these situations are usually set up with a curb separating left turning and through traffic on the main road.  While that is a good safety practice, I don't think the curb is really necessary unless there is a left turn movement from the side street to the main street.  In that circumstance, you need the curb to provide a channelization function so that the side street left turn traffic does not turn dirctly onto the main lanes that see a perma-green (or equivalent).  The left turners in those situations are to merge in somewaht later, as shown most clearly in US 89's third link.  Where side street traffic is not allowed to turn left, the curb is not necessary.  In situations where the side street is one-way away from the intersection (like in the circumstances involving on-ramps to freeways), I see no benefit at all to the separating curb.  But in situations where only right turns are allowed, but left turns are possible but illegal, the curb provides one more level of protection against an errant driver.  So this is a long way of saying that in the Rochester situation, the curb is not necessary because traffic from the shopping center must turn right, but a curb will probably be a good idea to prevent illegal lefts that may be very tempting at this intersection.

[mrsman]

Found a two section signal head still in use in Brooklyn, NY.
https://goo.gl/maps/3HF5vyCes9vZMEqM6

The funny thing is it's only for one direction. The others have the normal three section heads.

It's a 2 head red blinker, although looking at a series of GSVs coming down the street, it may be a steady red at top for a time period, then alternating red on top & bottom. https://goo.gl/maps/sYuQfvFNar5Mb3L89 then https://goo.gl/maps/JeNHQbBdefvYkvnt9

Those 2 section red lights are usually solid red on top, then change to flashing red on the bottom. Usually used for a driveway or other special circumstance.

These are quite unique signals.  Basically as it toggles between a solid red and a flashing red.  During solid red, it's the equivalent of a red light, but during flashing red it acts like a stop sign.

When signed at a driveway, it imposes a stop control during the flahsing phase.  The most practical application of this is that it means that traffic coming out of the driveway has to yield to opposing left turns, which is contrary to the usual practice.  For a really minor driveway, it is OK, but I don't think it fits for many busy driveways, like out of shopping centers.  IMO, it's better to sign a full RYG signal instead.

In the situation above, the signal is used at one of the many service roads that can be found on wide NYC streets.  The service roads are meant to access parking, driveways, and side streets, leaving the main road unimpeded.  The number of conflicts with this arrangement are staggering and it would not be safe unless the service road traffic is really quite small.  And putting stop signs at every cross street instead of signals helps discourage through traffic from the service roads.  (And this distiguishes the service roads on Ocean Parkway from the service roads on other streets like Kings Hwy or Queens Blvd.  The other streets have a lot of traffic on the service roads and the service roads are fully signalized.)

So what we have at the corner of Neptune and Ocean Parkway is an enhanced stop sign at the service roads.  I presume the service road sees a solid red (no movement at all, even right turns) whenever parallel main road traffic has a solid red (the times when left turns or Neptune Ave traffic has the right of way) and gets the flashing red whenever the main road gets a green.  This allows for traffic to proceed, but being watchful for pedestrians (who very frequently jaywalk across the service road) and from possible turning traffic from the main road.  It seems that right turns from the main road are prohibited, but I don't think this was always the case.  If right turns from the main road are allowed, only stop signs (or flahsing red) would be appropriate so that the traffic on the service road properly yields to all other traffic in its path.

[traffic light guy]

I've noticed that in Southeast Pennsylvania, there are these weird looking Durasigs. Typically they're found in the vicinity of Abington Township, typically around Glenside, hence the nickname I gave them; "Glenside style" Durasigs.

They have aluminum Siemens visors unlike the standard ones. These were mostly installed sometime between the time that they discontinued the 2nd gen ones with the black ring around the lenses and the third generation ones.

Here are several intersections that have them:

  https://www.google.com/maps/place/New+Vision+Youth+%26+Community/@40.1042768,-75.1645017,3a,75y,84.45h,94.54t/data=!3m6!1e1!3m4!1sN0ENmg5URxJRdvzI5nH5XQ!2e0!7i16384!8i8192!4m13!1m7!3m6!1s0x89c6ba0ab2d8cf81:0xc03bc345ceabce59!2sEaston+Rd+%26+E+Waverly+Rd,+Glenside,+PA+19038!3b1!8m2!3d40.098642!4d-75.1554277!3m4!1s0x89c6ba71eae9e445:0x300eab4594312819!8m2!3d40.1062977!4d-75.1646793

https://www.google.com/maps/@40.1079437,-75.1635287,3a,75y,44.36h,94.65t/data=!3m6!1e1!3m4!1sBd7Fp-WTlJwmNE7LllRoNQ!2e0!7i16384!8i8192

https://www.google.com/maps/@40.1134253,-75.1526764,3a,75y,219.8h,107.71t/data=!3m7!1e1!3m5!1sNYvzUN-_1McYBW-8gAtc8Q!2e0!6shttps:%2F%2Fstreetviewpixels-pa.googleapis.com%2Fv1%2Fthumbnail%3Fpanoid%3DNYvzUN-_1McYBW-8gAtc8Q%26cb_client%3Dmaps_sv.tactile.gps%26w%3D203%26h%3D100%26yaw%3D122.48542%26pitch%3D0%26thumbfov%3D100!7i16384!8i8192

https://www.google.com/maps/@40.1326352,-75.1528154,3a,75y,168.98h,91.66t/data=!3m7!1e1!3m5!1sCu1fBq_w1vQ_Do42Jtgt_A!2e0!6shttps:%2F%2Fstreetviewpixels-pa.googleapis.com%2Fv1%2Fthumbnail%3Fpanoid%3DCu1fBq_w1vQ_Do42Jtgt_A%26cb_client%3Dmaps_sv.tactile.gps%26w%3D203%26h%3D100%26yaw%3D75.68352%26pitch%3D0%26thumbfov%3D100!7i16384!8i8192

[mrsman]

Here is a very good article on the pedestrian buttons at traffic signals that pedestrain activists derisively call beg buttons.  It provides the right type of defense and really explains their function.

https://ggwash.org/view/86231/an-engineers-defense-of-the-pedestrian-push-button


Of course the activist commenters ridicule the responses of the traffic engineer because they don't seem to mind when drivers are forced to wait for nobody, which is all too common in non-actuated signals.  I happily choose to ignore them.

[US 89]

There was a midblock crosswalk I used to use regularly in Atlanta that did not have buttons at all - rather, it had pedestrian sensors that were supposed to detect if there were people waiting to cross and change the light accordingly.

The sensors sucked. Often times they simply wouldn't detect pedestrians at all until enough of them showed up, but other times they'd change the light when nobody was there. And as a pedestrian, there was no way to know if the light was just waiting for the right timing to change or if the sensors simply hadn't seen you.

They wound up getting rid of the sensors entirely and putting in buttons. That should be the way to go.

[Amtrakprod]

Here is a very good article on the pedestrian buttons at traffic signals that pedestrain activists derisively call beg buttons.  It provides the right type of defense and really explains their function.

https://ggwash.org/view/86231/an-engineers-defense-of-the-pedestrian-push-button


Of course the activist commenters ridicule the responses of the traffic engineer because they don't seem to mind when drivers are forced to wait for nobody, which is all too common in non-actuated signals.  I happily choose to ignore them.

I always tell them, you're not mad at the button, you're mad at the signal timing.


iPhone

[Amtrakprod]

There was a midblock crosswalk I used to use regularly in Atlanta that did not have buttons at all - rather, it had pedestrian sensors that were supposed to detect if there were people waiting to cross and change the light accordingly.

The sensors sucked. Often times they simply wouldn't detect pedestrians at all until enough of them showed up, but other times they'd change the light when nobody was there. And as a pedestrian, there was no way to know if the light was just waiting for the right timing to change or if the sensors simply hadn't seen you.

They wound up getting rid of the sensors entirely and putting in buttons. That should be the way to go.

Just because one experience with Ped detection was bad doesn't mean they all are. We've had good experience with Ped detection, it just needs a back up button and confirmation. Also, many signals should have Ped phases on recall.


iPhone

[jakeroot]

I used to walk to school every day for a couple of years. The street I used (Market St in Tacoma, WA) was entirely timed, alongside most of the intersections in the area. Most of the time, I jaywalked. I had no more of an interest in waiting for a walk signal than the car next to me waiting for their green. Market, alongside every other road in the area, was not timed for pedestrians, obviously. Ergo, there was no "WALK"  band afforded to me. Actuated signals with push buttons would've save me lots of time, as I was usually waiting for no one, and I could have changed the signal within a few seconds of arriving.

I have only one issue with pedestrian push buttons: if you arrive too late, you have to wait a whole cycle before getting a walk sign again. This happens for very obvious reasons: if one direction gets a green while someone is already waiting at the cross street, they have the next priority. The controller has already designated them for the next phase, and the current through phase will end as soon as vehicles stop arriving or it times out. There really needs to be a good four to six seconds at the beginning of a through phase to allow pedestrians to hit the button before it stops them from receiving a WALK sign.

Now, this is where automatic WALK is actually very important. Corridors with so much traffic that it's effectively a guarantee that the through phase will last at least as long as the WALK sign...those movements should have automatic WALK.

Pedestrian detection methods are a whole other issue. As are, I assume, bike detection, as there is already robust methods that don't require any button pressing.

[Scott5114]

There was a midblock crosswalk I used to use regularly in Atlanta that did not have buttons at all - rather, it had pedestrian sensors that were supposed to detect if there were people waiting to cross and change the light accordingly.

The sensors sucked. Often times they simply wouldn't detect pedestrians at all until enough of them showed up, but other times they'd change the light when nobody was there. And as a pedestrian, there was no way to know if the light was just waiting for the right timing to change or if the sensors simply hadn't seen you.

They wound up getting rid of the sensors entirely and putting in buttons. That should be the way to go.

Just because one experience with Ped detection was bad doesn't mean they all are. We've had good experience with Ped detection, it just needs a back up button and confirmation. Also, many signals should have Ped phases on recall.

Speaking as someone who writes code here and there...accurate pedestrian detection is such a difficult problem to solve that it's unlikely that a really good implementation of it can be done for the price that a government would want to pay for it.

Or as Past Scott said once...

Vehicles are a lot easier to detect than pedestrians, simply because vehicles are a lot more uniform in their composition, construction, and behavior than humans are. Cars nearly always contain enough iron to be magnetically detectable. They nearly always appear in the lines between lanes, so they can be detectable visually. Meanwhile, pedestrians can be anywhere from 3 to 7 feet tall and weigh from 50 to 400 pounds and rarely travel in straight lines and wait in uniform places. They're not always shaped the same (some are missing parts, use things like wheelchairs, walkers, or canes, or are carrying things like grocery bags that distort the silhouette). And if your ped-detecting algorithm happens to choke on a certain type of person that happens to coincide with a protected class (like, say, a dark background making it so dark-skinned people aren't reliably detected), now you've got a civil suit you've got to deal with.

Ideally, whenever a button is used, it wouldn't matter where in the cycle the light is, the button would trigger the requested pedestrian cycle more or less immediately (which seems to be the case in Norman whenever I've used a ped button). Even if you had flawless pedestrian detection equipment, it would be a scant improvement over a button if the request isn't serviced in a timely manner, which I suspect is your real beef with button-actuated signals.

[Lukeisroads]

do you think this signal is useless it leads to a service center in buena park https://www.google.com/maps/@33.8703489,-118.0031917,3a,75y,302.11h,89.95t,0.92r/data=!3m6!1e1!3m4!1sh7995nQ4hQb4o8CW8lZMpQ!2e0!7i16384!8i8192

[US 89]

do you think this signal is useless it leads to a service center in buena park https://www.google.com/maps/@33.8703489,-118.0031917,3a,75y,302.11h,89.95t,0.92r/data=!3m6!1e1!3m4!1sh7995nQ4hQb4o8CW8lZMpQ!2e0!7i16384!8i8192

I will say I don't think I've ever seen Driveway abbreviated as "Dwy" on a street blade before...

[jakeroot]

The fact that it's a fully protected left says so much about California's signal practices.

[Amtrakprod]

The fact that it's a fully protected left says so much about California's signal practices.

Looks like a good candidate to me, tbh.


iPhone