New Idiom in Traffic Signaling

The City of Portland is trying a new approach to “left turn yield to oncoming traffic”:

City of Portland installs new experimental Left-Turn Signal with Flashing Yellow Arrow.

The City of Portland’s Office of Transportation will be activating a new flashing yellow arrow left-turn display at NE 82nd and Tillamook on Tuesday, June 6th at 10:00 AM. Previously, drivers could only turn left on a green arrow like at many other intersections in Portland. Often drivers are annoyed when they see gaps in opposing traffic, but are not able to turn left because of a red arrow. On 82nd Avenue a new flashing yellow arrow will be displayed during the time that opposing through traffic has a green light and drivers will be permitted to turn left after yielding to oncoming traffic and pedestrians. Left turners would still get a green arrow if they are not able turn during the flashing yellow arrow.

“This new flashing yellow indication will reduce delay and frustration for Portland drivers” said Commissioner Sam Adams. Adams continued “From a safety perspective, drivers should make sure they have enough time to turn safely during the flashing yellow arrow. If the gaps are not long enough, then they will eventually get a green arrow.”

The new signal display will consist of a combined steady green arrow / flashing yellow arrow, a steady yellow arrow and a steady red arrow. The green arrow indicates that the left-turning traffic has the right-of-way, and the flashing yellow arrow tells left-turning traffic to first yield to oncoming traffic before turning.

Other future Portland test sites will be at SW Barbur Boulevard and 30th Avenue, and at NE 39th Avenue and I-84. If the flashing yellow arrow improves motorist safety and reduces traffic delay at the test sites, then there may be other locations in Portland where this new type of traffic signal might be installed.

Discussion on some neighborhood list servs is that this could be less safe for pedestrians. Thoughts?


21 responses to “New Idiom in Traffic Signaling”

  1. I don’t understand the need to “test” this. Thousands of towns/cities all over the country already have these, or don’t have a dedicated turn signal at all. I’m sure plenty of data exists that would be useful in determining accident rates, etc. It’s not rocket science – review the data and make a decision already.

  2. Left turners would still get a green arrow if they are not able turn during the flashing yellow arrow.

    While I support the idea of drivers being able to turn left without waiting for a dedicating green light –I am a driver, after all– this is additional time that pedestrians are put at risk when they are walking with their green light.

    The reality is that drivers are way more apt to be looking for cars coming the other direction, especially on a busy street, then to be looking for pedestrians coming up to the crosswalk.

    For example, on SW 5th and Clay downtown, where drivers turn on the red light, they’re looking to their left for on-coming vehicles, not looking to the right to see if there are pedestrians in the crosswalk they’re about to cross. This led, some 7 or 8 years ago, to a Tri-Met bus driver hitting, dragging, and killing a young woman who was crossing in the crosswalk with the light. (Since I crossed here often at the time, I complained a number of times to Tri-Met before the accident, to no avail.)

    I guess what I’m suggesting is that by eliminating un-necessary delays for cars, we’re also decreasing the gaps in traffic for pedestrians to cross the street. Pedestrians never get a dedicated green light.

  3. So it seems to me the comparison to be made is with an ordinary intersection that allows all turning movements on green.

    In that case pedestrians have a walk signal and cars can turn left, but have to yield for any pedestrians.

    I’m assuming that on a yellow arrow, pedestrians also have a walk signal. Do we think cars are less likely to yield to peds on a yellow arrown than they would on a green?

  4. Chris –

    I’m not ready to assume that pedestrians have the walk signal at the same time as there is a yellow arrow… it may be a “push button to cross” situation, where when the pedestrian has a walk signal, the left-turn signal is red only.

    I want to see it in action first.

    By the way, I used to be very opposed to “press button for walk signal” setups, as I felt they were very unfriendly to pedestrians. But, two things have changed my thinking and I now believe they are OK in limited circumstances.

    1. Newer push-buttons provide visual and sometimes audible feedback that the button has been pushed. When I am a pedestrian, I am much more comfortable waiting for the light to cycle when I am confident that my position in the queue has been registered. Otherwise, you often see people pushing the button repeatedly and giving up in frustration, jaywalking because they are unsure if they will ever get a real chance to cross.

    2. During the time I lived in Corvallis, I saw a very innovative traffic signal configuration installed (at 15th and Western) about 10 years ago. At night and during periods of light traffic, the signal is red in ALL directions. If you are the only car approaching the signal, it will turn green for you almost instantly, but you have to be going slow enough, of course, that you could stop for the red if the light does not change. The net result was traffic calming… it did not pay to speed along the road at night, but if you were driving slowly and smoothly enough, you were rewarded with a quick change to green without having to come to a full stop. In order to have a signal that can react this quickly, however, pedestrians had to push a button to be recognized. On the plus side, if you were the only person there, you would get the walk signal instantly.

    – Bob R.

  5. So it seems to me the comparison to be made is with an ordinary intersection that allows all turning movements on green.

    Maybe, but I think the better comparison is with an intersection where the driver can’t turn whenever they want, but has to wait for the green turn light. Giving just a little bit of extra breathing room for the pedestrian to only have to deal with cars turning from one direction, not both.

    I don’t think its that big a deal, but what struck me (so to speak) about the announcement was how it talks about how the test will measure whether it “improves motorist safety and reduces traffic delay at the test sites” and mentions nothing about what impact it may, or may not, have on pedestrians. One of the test sites will be on NE 39th, and I still drive by and think about –on SE 39th– the poor young woman killed not long ago in the crosswalk when a driver turned off 39th right into her.

    Part of what makes the 82nds and 39ths a little different than the average intersection is the cars –legally or not– are traveling faster, and are more distracted from the traffic, shops, buses and whatever.

  6. Reacting to the last two comments:

    1) I think PDOT pretty much has a policy of avoiding push-to-walk signals unless there are no good alternatives.

    2) The discussion reminds me of the argument that the best result is by removing all signalization/signing, forcing people to pay attention to what’s going on!

  7. “Walk signals” do not come on during green left turn signals. Our schedulers are smarter than that.

  8. Ron, I wasn’t suggesting that they did. What I was saying is that in a simple two phase signal, when green allows movements straight, right or left, motors can turn when pedestrians have a walk signal but must yield to pedestrians.

  9. Ron –

    I know that “walk signals” do not come on during green left turn signals. The question was whether, with this new light, will walk signals come on during the special flashing yellow left turn condition?

    By the way, there are some interesting exceptions out there right now which I consider dangerous to pedestrians. For example, there is a signal at 58th and Glisan.

    At this signal, “58th” is really just the offramp from I-84 eastbound. Coming from this ramp, your only choice is to turn left or to turn right onto Glisan. The lane markings include the kinds of arrows you would find in a left-turn-only lane at a major intersection.

    However, the traffic signals are solid green circles, not arrows, and when they are green and traffic exiting the freeway is coming up the ramp and turning left or right, the pedestrians also have a “walk” signal that puts them right into harm’s way.

    Additionally, at this very same intersection, there is a dedicated right-turn lane from westbound Glisan onto the I-84 ramp. There is signal with a sign that says “right turn signal”. A driver has a reasonable expectation that this is treated as a dedicated signal phase such as the usual left-turn signal, but it is not! It conflicts with pedestrians walking along Glisan, passing in front of the freeway ramp.

    This is a push-button-for-walk intersection (IIRC). You would think that once the pedestrian has a signal phase, that it would be a protected time to cross without interference from high-speed left/right turns, but it is NOT!

    I live near this particular intersection, and I am ashamed to say that I almost hit a pedestrian who had a walk signal at the same time I had a green-arrow “right turn signal”. Scary stuff.

    – Bob R.

  10. Not to muddy the issue too much, but how might the pedestrian scramble (all-pedestrian light cycle period) fit into any of the above scenarios?

    For a while now, I’ve been of the belief that downtown Portland should re-time all of their traffic signals. I see two options:

    One: Green lights would be for cars only, to allow for faster clearing of intersections and turn traffic. However, the green period would be shortened slightly, and a pedestrian scramble cycle inserted, during which no automobile movement would be allowed, but pedestrians could cross any direction they pleased.

    Two: half of every green light period would give the OK to pedestrians. The other half would give a RED hand to pedestrians, to allow for turning movements of cars. There would then be an additional cycle, the pedestrian scramble, during which no automobile movement would be allowed, but pedestrians could cross any direction they pleased.

    Perhaps downtown might not be the best first place to try this, (Broadway/Weidler or another basically stand-alone couplet might be more appropriate), but the reasons that I suggest it are these:

    1. Pedestrian scrambles would seem to be the safest way to get pedestrians across the street. There are no turns on red allowed during the scramble cycle, so there is no potential for ped/vehicle conflict as long as everybody plays by the rules.
    2. Initiating the scramble cycle at every intersection in a grid would allow for the entire grid to be re-timed at once. This would mean that, as you were driving along, you would still get greens as you progressed from one intersection to the next (assuming, of course, that you drove at the proper speed and congestion was not an issue). However, following behind the green cycle for cars would be an all-red cycle for cars — the scramble cycle, when pedestrians would go for it.

    I think there are two potential variations on this theme:

    a) The lights cycle green for cars heading north/south, green for cars heading east/west, then the pedestrian scramble cycle.

    b) The lights cycle green for cars heading north/south, then pedestrian scramble, then green for cars heading east west, then a second pedestrian scramble cycle!!

    I’ve only ever seen the first option used in the real world. I’d love to see the second option tested in a high-pedestrian-volume area like downtown Portland!

    Finally… bicycles. As I see it, bikes could go on either the appropriate green cycle for cars, or, if they were careful and slowed down, they could also traverse the intersection during the pedestrian cycle, if they did so at pedestrian speeds.

    If pedestrian scrambles became the norm at every signalized intersection, then I think the blinking yellow light left-turning issue would become a moot pedestrian safety consideration. I think it’s a great idea to reduce motorist dwell time and reduce congestion at left-turn signals.

    Barring the pedestrian scramble, however, Frank does raise a good point about the added potential for ped/vehicle conflict.

  11. Three Oregon entities; ODOT, City of Beaverton and Jackson County participated in a NCHRP study on flashing yellow turn signals. You can find the various working papers and reports at the link below.

    NCHRP Report

  12. Garlynn, about 10+ years ago I experienced the “pedestrian scramble” signals on a trip through parts of southern Oregon. I think it was either in Roseburg or Klamath Falls, can’t remember which. The all-pedestrian signal phase included the opportunity to cross the street diagonally, which saves the pedestrian a bit of time if they need to cross two sides of the same intersection.

    – Bob R.

  13. Bob-

    Yes, the diagonal crossing feature is my personal favorite aspect of the scramble, as well. I find it rather liberating to be able to cross an intersection diagonally, and legally, while the cars wait. It’s quite a feeling to have that pedestrian-only time in the intersection!

    Would you say that you felt a preference for option a or option b (green, green, scramble, or green, scramble, green, scramble), as given above, based on your real-life experience with it?

  14. Well, speaking as someone who believes that the pedestrian environment should be improved, but also as someone who drives and appreciates niceties like right-turn-on-red that would go away in a pedestrian scramble, I can’t say that I really prefer either system at this time.

    I would be willing to consider Option A for areas with consistently large numbers of pedestrians, such as around a school or popular facility. Perhaps in the future, traffic lights will be smart enough to have sensors to determine pedestrian demand as well as automobile demand… Imagine a sports arena or performance hall where the signals in the area could detect a sudden increase in pedestrian activity and automatically switch to “pedestrian scramble” mode each cycle until the demand peak has passed.

    As for general use in the downtown grid, I think it would throw the timing way off. I read somewhere that the downtown grid timing is about 60 seconds per cycle and could be practically brought down to 54 seconds if needed. To introduce the extra scramble phase would then extend the cycle time to 81 seconds (27 sec X 3 phases).

    The cycle time is a primary limit on the number of buses and light rail trains that can pass through the grid per hour. By increasing the cycle time about 40%, you correspondingly reduce peak capacity.

    (The reason cycle time is a limiting factor is because at every regular bus stop or train station, while the train is loading/unloading, it will lose the green light and must wait for the cycle to come around again before the train can proceed.)

    This would happen to an eastbound MAX train, for example, at 3rd, 5th, Pioneer Square, and 10th, adding nearly 2 minutes to the total transit travel time through that segment, which is already the slowest stretch in the entire system.

    – Bob R.

  15. PS… I think 1st & Oak would also be included in this scenario, bringing us to about 2.4 minutes.

    – Bob R.

  16. Bob-

    That’s a good point about the light rail. Downtown Portland’s signal cycles are very delicately balanced right now, and for the light rail reason alone, it’s probably not a good idea to fix a system that isn’t broken…

    However, just hypothetically speaking, the reason why I suggested downtown Portland’s grid is because, for the scramble cycle to work anywhere in a grid, I believe it must be turned on everywhere in the grid. Otherwise, signal timing breaks down where the incongruity lies. In other worse, it would not work to introduce an 80-second cycle at 5th & Morrison, if every other intersection in the system continued to operate on a 60-second cycle. Again, hypothetically speaking, it would seem that either those 60 seconds could be broken into thirds, such that traffic in each direction got 20 seconds, and pedestrians their own 20 seconds, or the entire grid could move to an 80-second cycle. But the entire grid would need to change at once.

    Similarly, for any timed lights, the entire sequence would need to change at once. Broadway & 12th, for instance, could not institute a scramble unless the entire Broadway/Weidler couplet moved to include the scramble cycle. Otherwise, chaos would ensue, and the intersection with the scramble cycle would end up screwing up the traffic flow on the rest of the couplet.

    I’ve actually seen this happen, which is why I bring it up.

    In Oakland, CA, a scramble cycle was instituted at 8th & Webster. However, Webster is timed, and is part of a grid. Traffic flowing down the grid towards 8th invariably backs up at the scramble intersection, as the rest of the lights were not programmed with a scramble cycle. The end result is a bloody mess on Webster, backing up from 8th through 9th towards 11th — all day long.

    However, if the scramble cycle were instituted at all the intersections, then the lights could resume being timed normally, with every intersection receiving a scramble cycle, regardless of pedestrian volume.

    Finally, addressing your right turn on red question… I think that needs to be examined. Right turns on red should continue to be OK, but perhaps a new rule, called “No Right Turn During Scramble” would help sort everything out. Or, perhaps no new rule is needed, as long as it is perfectly clear that right turns on red are permitted, but vehicles must yield to pedestrians, and if there is a pedestrian anywhere in an intersection during a scramble, no vehicle may enter the intersection until after the pedestrian has cleared it.


  17. Garlynn –

    I don’t think we could break it down into 20-second thirds. A minimum amount of time must be allowed for pedestrians (including slower moving elderly and disabled) to make it across the intersection, especially if we allow diagonal crossings which are inherently longer than direct crossings and put the pedestrian in front of two sets of traffic.

    I believe the 27 second figure I had heard was the practical minimum due to pedestrian considerations.

    Regarding right-turns on red, I have seen signals in parts of Wisconsin that used a flashing right arrow during the red cycle to indicate that turns were possible, but I do not remember the exact context of the other signal phases at the time.

    – Bob R.

  18. Bob-

    Agreed, 20 seconds is probably too little — which is why I said “hypothetical” in front of it. I threw it out there more as an illustration of the concept of ways to incorporate the pedestrian scramble into a timed-light grid. The grid must maintain consistency at all intersections to be effective.

    However, Broadway/Weidler could still be a potential test case for the concept… as well as MLK/Grand. Or both, since they cross at one point. The pedestrian volumes might not be there yet, but they could be in the future.

    Another place it may be useful is in NW Portland — NW 21st & 23rd at Glisan & Everett, for example, to the extent that those intersections could be timed to include scrambles without hosing up the rest of the lights in NW Portland.

    Of course, this may be a solution searching for a problem (this=the integration of the pedestrian scramble as a cycle in a timed-traffic-light couplet or grid system). Perhaps Portland just doesn’t have the pedestrian volumes anywhere to necessitate the scramble yet?

  19. Garlynn –

    I would think the best candidates for pedestrian scramble implementation would be two-way streets with moments of peak pedestrian activity (sports arenas, schools letting out, etc.), where the scramble mode could be activated by smart sensors, or perhaps a remote activation switch in a school or sports facility office.

    Two-way streets are inherently difficult to coordinate signals… you can time them to give favor to a particular direction at the expense of the other direction, but otherwise you can’t smoothly regulate the movement of traffic like you can with a one-way grid or couplet. Therefore, introducing an extra cycle for pedestrians at select intersections and times won’t significantly disrupt the flow of traffic. (Unless the road is already operating at capacity.)

    – Bob R.

  20. It’s not safe, I was walking today and just when I had the walking sign. A car turned left because it had the green arrow, the person did not see me that they could have hit me. I saw the car turning that I stoped in the middle of the street to let the car turned. Just when they turned, they noticed me crossing the streets. It’s really scary.

Leave a Reply

Your email address will not be published. Required fields are marked *