Leading vs. lagging PPLT’s

[tradephoric]

What type of permissive-protected left turn (PPLT) do you prefer: leading or lagging? 

According to a recent survey, 83% of PPLTs in the US are leading while 11% are lagging.

[jakeroot]

Lagging. Signal phases (IMO) should start with solid greens in both directions, and then the turns should go protected to filter out the waiting traffic (if there is any -- otherwise, the opposite direction continues to see a green orb until the protected turning phase ends). This would help cure the notion that entering an intersection, waiting to turn left, might cause someone to run the red (despite having already entered the intersection, legally, on green).

AFAIK, you'd have to use an FYA to have the specific control that I'm describing.

[Brandon]

What type of permissive-protected left turn (PPLT) do you prefer: leading or lagging? 

According to a recent survey, 83% of PPLTs in the US are leading while 11% are lagging.

Lagging.  I'd rather clear the turn lane out after and leave zero vehicles sitting in the intersection waiting when the signal turns red.

[Brandon]

Lagging. Signal phases (IMO) should start with solid greens in both directions, and then the turns should go protected to filter out the waiting traffic (if there is any -- otherwise, the opposite direction continues to see a green orb until the protected turning phase ends). This would help cure the notion that entering an intersection, waiting to turn left, might cause someone to run the red (despite having already entered the intersection, legally, on green).

AFAIK, you'd have to use an FYA to have the specific control that I'm describing.

Not necessarily.  It would work just as well with a doghouse or tower if both directions go green-yellow-red, then green arrow-yellow arrow at the same time.  If you want to let turning traffic move while the other side has a green ball, then yes, an FYA would be better.

[tradephoric]

Here is a before/after video of an intersection that was converted from a lagging left (pre-FYA) to a leading left (FYA).

https://www.google.com/maps/@42.7203631,-83.418381,3a,75y,348.83h,77.06t/data=!3m4!1e1!3m2!1stMrE7Ocuz8GS7Y-Nqy_k-A!2e0

As has been noted, leading lefts can lead to drivers getting "stuck" in the middle of the intersection at the end of the permissive phase.  When the side-street turns green, side-street traffic often has to wait for left-turning vehicles to clear (examples of this can be seen at 0:15, 2:15, and 3:15 in the FYA video).

With lagging lefts, the potential for the "yellow trap" increases significantly.  To prevent the yellow trap in the pre-FYA example, all through phases must terminate before the protected left comes up, which reduces intersection capacity.

[jakeroot]

Here is a before/after
With lagging lefts, the potential for the "yellow trap" increases significantly.  To prevent the yellow trap in the pre-FYA example, all through phases must terminate before the protected left comes up, which reduces intersection capacity.

I think that's what my brain was thinking earlier when I suggested that an FYA is the most desirable traffic control device for a lagging turn signal.

[dfwmapper]

Texas style. During peak hours, leading for the dominant direction of traffic flow and lagging for the other. At other times, based on the sensor, but leading for one or both depending on whether or not cars are waiting.

[tradephoric]

I think that's what my brain was thinking earlier when I suggested that an FYA is the most desirable traffic control device for a lagging turn signal.

In the video above, the opposing left has a LEFT TURN case sign.  As long as that case sign exists, lagging the left with a FYA would still lead to a "yellow trap".  This is a good example of how the FYA doesn't always solve the "yellow trap". 

Here's a view of the opposing left turn case sign.
https://www.google.com/maps/@42.7203631,-83.418381,3a,36.7y,338.48h,88.81t/data=!3m4!1e1!3m2!1stMrE7Ocuz8GS7Y-Nqy_k-A!2e0

[roadfro]

Texas style. During peak hours, leading for the dominant direction of traffic flow and lagging for the other.

When first done with 5-section displays (with a louvered circular green), this was called "Dallas Phasing" since that city did this type of phasing first on a widespread scale. I think many traffic engineering types still refer to this as Dallas Phasing, even though now you would use an FYA display instead.

[cl94]

Depends. Here are my thoughts:

Leading:
-Shared turn/through lane, only one direction gets protected phase
-Opposing protected lefts on busy highways (where, during much of the day, only a small amount of cars could turn left without a protected phase, in which case the turn should probably be protected-only)

Lagging:
-Dedicated turn lane, only one direction gets protected phase
-Opposing lefts where many vehicles in queue would be able to turn during daytime hours

Special cases/split phasing:
-If two opposing approaches have protected turn phases and no dedicated turn lane, one direction is leading and one is lagging (split phasing, all greens for approach with protected turn)
-If one direction has a shared left-through lane and the other has a dedicated turn lane and both have protected phases, approach without dedicated lane gets leading and other gets lagging. Yes, I have seen this and yes, it is going to be reconfigured next year.

Basically, as a transportation engineer, I'm for lagging unless there's a darn good reason not to use it, as it eliminates unnecessary phases, reduces lost time, and shortens the cycle length, all of which increase the LOS of the intersection.

In New York, Region 5 always uses leading unless there's some form of split phasing. Region 1, on the other hand, uses lagging (especially in Warren County) for turn movements without a protected left in the other direction (i.e. US 9 at I-87 Exit 20). I

[Pink Jazz]

Here in the Phoenix metro area, lagging left turn arrows are, with some exceptions, are the standard in the cities and towns of Scottsdale, Gilbert, and Goodyear.  Some other cities and towns also use lagging lefts at some freeway onramps, and the City of Tempe uses lagging lefts along Rural Road.

[roadfro]

Leading:
-Shared turn/through lane, only one direction gets protected phase

Unless you're running split phasing, you can't do protected left turns with a shared left/through lane...that will block through traffic.

Basically, as a transportation engineer, I'm for lagging unless there's a darn good reason not to use it, as it eliminates unnecessary phases, reduces lost time, and shortens the cycle length, all of which increase the LOS of the intersection.

One reason for leading PPLT would be physical constraints of the turn pocket. If you have a turn pocket that only stores 4 vehicles, but average demand is 6 left turning vehicles per cycle, using a leading phase can help clear out the turn pocket so that left turning vehicles are less likely to block the through lane waiting for the lagging left.


BTW: A lagging PPLT doesn't decrease cycle length, since the actual cycle time is generally fixed unless the signal is running free (in which case there is effectively no cycle)–a skipped phase doesn't change the cycle time, although an early phase release might give that appearance.

[cl94]

Leading:
-Shared turn/through lane, only one direction gets protected phase

Unless you're running split phasing, you can't do protected left turns with a shared left/through lane...that will block through traffic.

Basically, as a transportation engineer, I'm for lagging unless there's a darn good reason not to use it, as it eliminates unnecessary phases, reduces lost time, and shortens the cycle length, all of which increase the LOS of the intersection.

One reason for leading PPLT would be physical constraints of the turn pocket. If you have a turn pocket that only stores 4 vehicles, but average demand is 6 left turning vehicles per cycle, using a leading phase can help clear out the turn pocket so that left turning vehicles are less likely to block the through lane waiting for the lagging left.


BTW: A lagging PPLT doesn't decrease cycle length, since the actual cycle time is generally fixed unless the signal is running free (in which case there is effectively no cycle)–a skipped phase doesn't change the cycle time, although an early phase release might give that appearance.

I'd consider a turn pocket shorter than average demand a good reason to use leading left turns. By the first thing you quoted, it is a form of split phasing, but not pure split phasing. A skipped phase decreases the apparent cycle time in that there would be less waiting time for opposing phases. Cycle time really only applies in its "true form" if it's a purely timed signal and I was kind of trying to simplify things. Cycle time is a constant, but apparent cycle time is not if phases can be skipped.

[roadfro]

A skipped phase decreases the apparent cycle time in that there would be less waiting time for opposing phases. Cycle time really only applies in its "true form" if it's a purely timed signal and I was kind of trying to simplify things. Cycle time is a constant, but apparent cycle time is not if phases can be skipped.

Cycle time also applies in actuated-coordinated modes as well, but that beyond the simplistic view.

[froggie]

Leading:
-Shared turn/through lane, only one direction gets protected phase

Unless you're running split phasing, you can't do protected left turns with a shared left/through lane...that will block through traffic.

Not going to speak for cl94, but there are numerous examples in eastern states of PPLTs where there is only a single-lane approach to the intersection.  Fairly common on older, 2-lane roads where for whatever reason a left turn bay was not constructed when the signal was installed (whether due to cost, lack of ROW, or ROW impacts).  I've seen examples at both T-intersection and regular 4-way intersections.

[mrsman]

I would use lagging in all cases where a yellow trap or a perceived yellow trap would not occur.  Basically, if the opposing left (or u-turn) is prohibited, because of T-intersections and one-way streets, then a lagging left would not lead to a yellow trap.  Also, if the opposing left is protected only, a lagging left would also not lead to a yellow trap.

One great example of this is along Northern Blvd in Queens, NY.  Nearly every intersection has a signal and most streets are one-ways.  Since one direction is a prohibited turn, it is easy to add a permissive lagging left turn into the cycle.

http://goo.gl/maps/JZa3G

The great thing about a lagging left is that it allows for all the cars that waited a whole cycle to turn a chance at the end of the cycle.  And they don't have to worry as much about turning during the yellow phase.

It works so well, I would advocate adding a short lagging left turn phase to every 2-way/1-way intersection and every T-intersection, even if left turn volume is not high enough to otherwise justify the arrow.

However, in most intersections, the yellow trap is a real worry, so leading lefts must be used.

[cl94]

Leading:
-Shared turn/through lane, only one direction gets protected phase

Unless you're running split phasing, you can't do protected left turns with a shared left/through lane...that will block through traffic.

Not going to speak for cl94, but there are numerous examples in eastern states of PPLTs where there is only a single-lane approach to the intersection.  Fairly common on older, 2-lane roads where for whatever reason a left turn bay was not constructed when the signal was installed (whether due to cost, lack of ROW, or ROW impacts).  I've seen examples at both T-intersection and regular 4-way intersections.

That's what I was referring to. Also when there's a 2-lane approach, but no left turn lane for the same reason (lack of ROW, etc.). Very common in the east and increasingly so as one moves north toward PA, NY, and MA.

[roadfro]

Leading:
-Shared turn/through lane, only one direction gets protected phase

Unless you're running split phasing, you can't do protected left turns with a shared left/through lane...that will block through traffic.

Not going to speak for cl94, but there are numerous examples in eastern states of PPLTs where there is only a single-lane approach to the intersection.  Fairly common on older, 2-lane roads where for whatever reason a left turn bay was not constructed when the signal was installed (whether due to cost, lack of ROW, or ROW impacts).  I've seen examples at both T-intersection and regular 4-way intersections.

That's what I was referring to. Also when there's a 2-lane approach, but no left turn lane for the same reason (lack of ROW, etc.). Very common in the east and increasingly so as one moves north toward PA, NY, and MA.

Not something anybody in their right mind would intentionally design nowadays, but I'll concede the situation does exist.

I do recall seeing a situation like this in Washington DC about 6-7 years ago where a 2-lane approach had a protected left without a turn lane. (Couldn't tell you where this was, but ISTR traffic backed up coming off a bridge.) Nasty back up in that left lane since left turners waiting for a green arrow were blocking through traffic in the left lane, and the right lane was too congested to allow cars in the left lane to get over.

[tradephoric]

Basically, as a transportation engineer, I'm for lagging unless there's a darn good reason not to use it, as it eliminates unnecessary phases, reduces lost time, and shortens the cycle length, all of which increase the LOS of the intersection.

Leading lefts can be beneficial when you have uneven queues.  Suppose a NBLT averages 20 vehicles in queue and the SBLT has only 3 vehicles in queue.  With leading lefts you can allocate 40 seconds for the left turn phase and allow the SBLT to gap out early.   If the SBLT suddenly gets heavy, it can soak up the full 40 seconds if it needs it.  You couldn't do this with lagging left turns (you could set up the lagging lefts to start non-simultaneous and run the SBLT short, but then it wouldn't be able to service a sudden rush of SBLT traffic). 

For a standard 4-phase intersection, I personally believe leading lefts maximize intersection capacity but lagging lefts are safer (provided that the "yellow trap"  is eliminated).

[cl94]

Leading:
-Shared turn/through lane, only one direction gets protected phase

Unless you're running split phasing, you can't do protected left turns with a shared left/through lane...that will block through traffic.

Not going to speak for cl94, but there are numerous examples in eastern states of PPLTs where there is only a single-lane approach to the intersection.  Fairly common on older, 2-lane roads where for whatever reason a left turn bay was not constructed when the signal was installed (whether due to cost, lack of ROW, or ROW impacts).  I've seen examples at both T-intersection and regular 4-way intersections.

That's what I was referring to. Also when there's a 2-lane approach, but no left turn lane for the same reason (lack of ROW, etc.). Very common in the east and increasingly so as one moves north toward PA, NY, and MA.

Not something anybody in their right mind would intentionally design nowadays, but I'll concede the situation does exist.

I do recall seeing a situation like this in Washington DC about 6-7 years ago where a 2-lane approach had a protected left without a turn lane. (Couldn't tell you where this was, but ISTR traffic backed up coming off a bridge.) Nasty back up in that left lane since left turners waiting for a green arrow were blocking through traffic in the left lane, and the right lane was too congested to allow cars in the left lane to get over.

It certainly isn't ideal, but there just isn't the space to add a left turn bay in many locations due to development. Same reason you can't have Michigan lefts out here. Many municipalities up here are, for multilane highways, taking out a lane in each direction and adding center turn lanes and shoulders wide enough for on-street parking in at least one direction. May seem counter-intuitive, but traffic flows much more smoothly.

[lordsutch]

Leading if in permissive/protected mode, lagging if protected-only. That said I'm probably most used to leading lefts since it's what is commonly used in most of the places I drive on a semi-regular basis.

[tradephoric]

Reviving this thread from some years back! Here are SYNCHRO models comparing leading lefts vs lagging lefts at an isolated intersection with the exact same traffic volumes and splits.  (Volumes:  Lefts=400 VPH and Thrus=500 VPH; Split: Lefts=40 sec; Thrus=25 sec

https://youtu.be/qiS1PzGDVa0

https://youtu.be/fcU-9FpDDgU

The reason why the Lagging left model looks jammed up is because the left-turn phases are not capable of non-simultaneous gap out.  This leads to a lot of "wasted green" time.  Again, these left turn volumes are exactly the same but traffic on a cycle by cycle basis varies (one cycles the NBLT might be heavy while the next cycle it's the SBLT that's jamming).  Lagging-lefts are just not very good at dealing with varying traffic conditions.

[mrsman]

I also prefer lagging left turns since in most permissive left turns, there are always people who make the left at the end of the cycle.  Putting the left turn at the end of the cycle therefore seems natural to me.  Of course all concerns with yellow trap have to be met and the fya does that.

Leading lefts should maximize the capacity of an isolated intersection.  Suppose a NB LT gaps out early.  That extra time can be given to the SB THRU (even as the SB LT is still running).  With lagging lefts you can't reintroduce the THRU phase since it has already ran (so when that NB LT wants to gap out early it can't... and that's just wasted green time).

That may be true, but I've seen so many intersections where people wait for a gap to make a left turn.  As there usually is no gap, they end up making the turn once the light turns red and this delays cross street.  (And it's even worse if cross street has a leading left, as those are usually short phases.).  Not to mention that there is also a bit of a safety issue as they would have to make their turn against someone who could run a red light. 

A protected/permitted leading left turn, the left turners will make their turns at both the start and end of the cycle.  A protected/permitted lagging left will only make their turn at the end of the cycle (or when it is clear during the middle of the cycle without blocking other traffic).

Of course, every situation is different.  In CA, at least in the pre-FYA era, because of yellow trap concerns, lagging protected/permissive lefts are only allowed if the opposing left were prohibited or limited to a leading phase by "left turn on arrow only".  In such circumstances, they work great at clearing intersections with a left turn onto a one-way street.  They also work well, if the lagging left historically has many more cars than the leading left.  And, if one were concerned about pedestrians, leading protected/permissiv lefts are simply incompatible with the LPI.

So, in short, I agree that in most circumstances the leading left is preferred for the reasons that you mention.  [In my experience there are far more leading lefts for this reason.]  But if left turns against red light running, pedestrian safety, or where only one side of the intersection is making a left (or consistently a majority of the lefts), a lagging left would be preferred.


Another point:  The models that you created above seem to indicate protected only turning.  It seems that the left turners are only turning when they have a green arrow.  A common scenario with leading PPLTs is that one or two cars will turn at the end of the cycle, even if the light is already red and even if the road is very congested.  Look at the leading synchro at around 1:30, there is enough gap for NB cars to make the left, yet none are going because they face the left red arrow.

OK, in both protected-only and PPLT, your point stands - if the two left turns are imbalanced, if both directions are given a leading left, and NB gaps early, we can start SB green while still providing service to the SB left turn.  In a lagging turn, this is not a possibility.  And I think that if both directions were protected-only, I would also prefer that they both be leading for that reason to provide for the flexibility, and yes to avoid overall backups.  But I do not believe that this will hold true for PPLT.

Assuming that thru traffic is the heavier flow than left turning traffic, the goal to avoid overall intersection delay is to give left turn a protected green for as little time as necessary.  In some places, they won't even trigger a left turn arrow unless there are at least three cars waiting to turn left.  Now, if you have leading PPLT, you will trigger the left arrow as you have people waiting.  But it may not be necessary because if traffic flow is moderate, there could be gaps during the thru phase to make those lefts.  With lagging PPLT, the signal will know.  If the left turners are able to all make their left turns during the thru phase, the lagging left will not even be triggered for that cycle -saving everyone time.

[kphoger]

I've seen so many intersections where people wait for a gap to make a left turn.  As there usually is no gap, they end up making the turn once the light turns red and this delays cross street. 

To be fair, the cross street is typically only delayed by a couple of seconds when this happens.  A lot of times (in my experience), oncoming traffic is more willing to stop at a yellow light if they see people waiting to turn left and, as such, hardly any red time is used for left turns.  And, even when that's not the case, most of the left-turning happens during the all-red phase.

[paulthemapguy]

Leading.  Yellow trap is a phenomenon I really don't like, and a consistent use of leading left turns eradicates yellow trap from your system.   Though this Texas approach sounds like a cool idea now that I'm reading about it.