Are diverging diamonds a fad?

[jeffandnicole]

Yawn. DDIs are not intended for interchanges where the majority of surface traffic does not enter or exit the freeway. Therefore signal progression has no meaning when considering a DDI.

There are hundreds of SPUIs and Parclo A4s where good progression along the corridor should be a key concern.  However, these designs also stop both directions of travel and suffer from poor signal progression.  I know this thread was originally about DDIs but it has morphed into interchange designs that lead to poor signal progression (SPUIs and Parclo A4s included).  The point you are making in regards to DDIs can't be applied to the hundreds of SPUIs and Parclo A4s out there.   

Don't you want to drive 10, 15, or 20 miles along a major corridor and never get stopped at a red light?  Good progression is lost when progression killing interchange designs are chosen (ie. DDIs, SPUIs, & PARCLO A4s).

Timing lights to hit nothing but green along major corridors of these lengths is fairly unrealistic.  In a rural area, there's going to be too many varying speed differences to program the lights to offset from one another.  Even if you program the lights for the speed limit, chances are people will go above or below that limit.  In a suburban or urban setting, there's too many people entering and exiting driveways, parking lots, etc.  And eventually you're going to come along an opposing major corridor which is going to require a different light sequence than the corridor you're on.

If you can get a few miles of good signal coordination, I'll be happy with that.  10 miles with a few reds isn't too bad.  But to expect to go 20 miles with nothing but green lights is a near impossible feat.

[tradephoric]

Timing lights to hit nothing but green along major corridors of these lengths is fairly unrealistic.  In a rural area, there's going to be too many varying speed differences to program the lights to offset from one another.  Even if you program the lights for the speed limit, chances are people will go above or below that limit.  In a suburban or urban setting, there's too many people entering and exiting driveways, parking lots, etc.  And eventually you're going to come along an opposing major corridor which is going to require a different light sequence than the corridor you're on.

There have already been 3 videos posted in this thread demonstrating good two-way signal progression.  All of these corridors are found in Metro Detroit.  You won't see any progression killing DDIs, SPUIs, or Parclo A4s in these videos...

https://www.youtube.com/watch?v=eB2H4bGp4Jc
https://www.youtube.com/watch?v=7hvUm9vYJYI
https://www.youtube.com/watch?v=oea5WLUmGjs

If you can get a few miles of good signal coordination, I'll be happy with that.  10 miles with a few reds isn't too bad.  But to expect to go 20 miles with nothing but green lights is a near impossible feat.

Traveling 20 miles on US-24 without hitting a red light.  There is a caveat though as only the first 10 miles is capable of good two-way signal progression.
https://www.youtube.com/watch?v=Wkf7iLSblZ0

[briantroutman]

Don't you want to drive 10, 15, or 20 miles along a major corridor and never get stopped at a red light?

That would be nice, but lights are rarely timed for bike speeds.

This. You need a one-way grid for it to work, and even in downtown Portland, the signal timing leaves a bit to be desired.

Getting off topic, but here is downtown Portland's grid in action.  The city has very condensed grid so the 'green wave' that achieves perfect progression is only 12 mph (at 56 second cycle lengths).  A driver is pretty much limited to driving 12 mph through downtown Portland.

On most days, I cycle more miles than I drive. Under both circumstances in a dense urban district, I would much prefer to move peacefully along at a steady 12 or 15 m.p.h. than be constantly starting and stopping.

[vtk]

There have already been 3 videos posted in this thread demonstrating good two-way signal progression.  All of these corridors are found in Metro Detroit. 

Made possible by the fact that Detroit's arterials are nearly all mile-grid roads with Michigan Lefts.  How much of Boston or Atlanta would have to be demolished to make room for such a neat road system that is so agreeable to signal progression?

PS – I can imagine in my head some time/distance diagrams demonstrating SPUIs and A4s that don't ruin signal progression in a Detroit-like arterial grid. It wouldn't be perfect, and it would be highly dependent on the geometry being a certain way, (which is already a requirement for the kind of progression that gets tradephoric off), but it's possible. If I feel like sacrificing some of my free time for the task, I may share these diagrams in the next couple of days.

[tradephoric]

Made possible by the fact that Detroit's arterials are nearly all mile-grid roads with Michigan Lefts.  How much of Boston or Atlanta would have to be demolished to make room for such a neat road system that is so agreeable to signal progression?

If signal progression is already bad, like in the case of Boston or Atlanta, wouldn't that be a good reason to choose interchange designs that improve upon progression?  If someone is drowning would you try to hold their head under the water?

[vtk]

Made possible by the fact that Detroit's arterials are nearly all mile-grid roads with Michigan Lefts.  How much of Boston or Atlanta would have to be demolished to make room for such a neat road system that is so agreeable to signal progression?

If signal progression is already bad, like in the case of Boston or Atlanta, wouldn't that be a good reason to choose interchange designs that improve upon progression?  If someone is drowning would you try to hold their head under the water?

No, it doesn't work like that, and you missed my point. Unless you completely reconfigure the arterial network on a large scale in cities like that, the best signal progression that's even possible is still hardly recognizable as any progression at all. Sure, a choice between interchange designs may have an impact on signal progression, but we're probably talking about two nearly indistinguishable shades of crap, compared to Metro Detroit.

If someone's drowning, do you fuss over whether it's Lake Erie water or Evian that they are drowning in?

[cl94]

Made possible by the fact that Detroit's arterials are nearly all mile-grid roads with Michigan Lefts.  How much of Boston or Atlanta would have to be demolished to make room for such a neat road system that is so agreeable to signal progression?

If signal progression is already bad, like in the case of Boston or Atlanta, wouldn't that be a good reason to choose interchange designs that improve upon progression?  If someone is drowning would you try to hold their head under the water?

No, it doesn't work like that, and you missed my point. Unless you completely reconfigure the arterial network on a large scale in cities like that, the best signal progression that's even possible is still hardly recognizable as any progression at all. Sure, a choice between interchange designs may have an impact on signal progression, but we're probably talking about two nearly indistinguishable shades of crap, compared to Metro Detroit.

If someone's drowning, do you fuss over whether it's Lake Erie water or Evian that they are drowning in?

Thank you. Interchange design in much of the world is designed to get people onto the limited-access road as fast as possible, thus removing traffic from surface streets. This is the best design if arterials are clogged and the freeway has capacity, as you're minimizing traffic that backs up on the surface road.

What works in one place doesn't necessarily work everywhere. Understand that.

[jakeroot]

Interchange design in much of the world is designed to get people onto the limited-access road as fast as possible, thus removing traffic from surface streets. This is the best design if arterials are clogged and the freeway has capacity, as you're minimizing traffic that backs up on the surface road.

But weren't freeways sort of designed as bypass roads? If you introduce too much local traffic onto them, you get congestion on the freeway.

I understand how people often use freeways, but to me, they should be used to go from A to C by bypassing B, not to go from A to B to C (in the former, you can still access C, but you must exit at an arterial and drive the extra distance).

[tradephoric]

Made possible by the fact that Detroit's arterials are nearly all mile-grid roads with Michigan Lefts.  How much of Boston or Atlanta would have to be demolished to make room for such a neat road system that is so agreeable to signal progression?

If signal progression is already bad, like in the case of Boston or Atlanta, wouldn't that be a good reason to choose interchange designs that improve upon progression?  If someone is drowning would you try to hold their head under the water?

No, it doesn't work like that, and you missed my point. Unless you completely reconfigure the arterial network on a large scale in cities like that, the best signal progression that's even possible is still hardly recognizable as any progression at all. Sure, a choice between interchange designs may have an impact on signal progression, but we're probably talking about two nearly indistinguishable shades of crap, compared to Metro Detroit.

If someone's drowning, do you fuss over whether it's Lake Erie water or Evian that they are drowning in?

What are you arguing here?  I would prefer to drive on a Boston corridor where 90% of the signals have crappy progression as opposed to a Boston corridor where 100% of the signals have crappy progression.  The fact is any city could benefit from interchange designs that improve upon progression.

[cl94]

Interchange design in much of the world is designed to get people onto the limited-access road as fast as possible, thus removing traffic from surface streets. This is the best design if arterials are clogged and the freeway has capacity, as you're minimizing traffic that backs up on the surface road.

But weren't freeways sort of designed as bypass roads? If you introduce too much local traffic onto them, you get congestion on the freeway.

I understand how people often use freeways, but to me, they should be used to go from A to C by bypassing B, not to go from A to B to C (in the former, you can still access C, but you must exit at an arterial and drive the extra distance).

Eh, not always. Take New York City's expressway system. Really was never meant to do anything than get cars off of the surface roads.

[vdeane]

In some jurisdictions, the emphasis is on optimizing the performance of each intersection/interchange, not on having drivers hit a stream of green lights.  This is especially true for areas that primarily use actuated signals rather than fixed timed ones.  Many NYSDOT signals don't have a pre-programed timed sequence during ANY part of the day, not even rush hour.

[6a]

In some jurisdictions, the emphasis is on optimizing the performance of each intersection/interchange, not on having drivers hit a stream of green lights.  This is especially true for areas that primarily use actuated signals rather than fixed timed ones.  Many NYSDOT signals don't have a pre-programed timed sequence during ANY part of the day, not even rush hour.

How about a preference for trucks exiting a rail yard, and entering an interstate? 

The DDI here has simplified my commute immensely.

[tradephoric]

Are the NYSDOT actuated signals mainly along rural high speed arterials where the signals are more spread out? 

[Scott5114]

Signal progression is all well and good but interchanges often have concerns that are of more importance.

Of course, I live in a part of the country that just got their first SPUI two years ago and still thinks left exits are pretty neat, so I doubt signal progression is a realistic thing to hope for. OK-9 in Norman would be a great candidate for it, but no such luck.

[cl94]

Are the NYSDOT actuated signals mainly along rural high speed arterials where the signals are more spread out?

Not necessarily. Near Buffalo, almost every NYSDOT signal is at least semi-actuated, including in dense parts of the suburbs. Traffic patterns are too erratic to do simple timing and the advanced demand-based systems are very costly. Only real exceptions are the few NYSDOT signals in downtown Buffalo, which are on green waves. Buffalo maintains most of the signals in the city, even along the few state-maintained roads (excluding Elm/Oak and Church, which are the aforementioned timed NYSDOT signals). There's a project going on to coordinate the signals along NY 78 and NY 324, but that's really it for state stuff. This holds true in most regions. NYC is synced, but every signal in the City is city-maintained.

The I-270 example is pretty similar to how preference is assigned in most places. Might ruin any chance of progression, but it's very effective at getting trucks between the intermodal yard and the expressway.

[kphoger]

Signal progression is all well and good but interchanges often have concerns that are of more importance.

This! I thought it was commonsense, but apparently not.

[tradephoric]

Signal progression is all well and good but interchanges often have concerns that are of more importance.

This! I thought it was commonsense, but apparently not.

Here is a comparison analysis of the diverging diamond interchange and partial cloverleaf interchanges:

http://fau.digital.flvc.org/islandora/object/fau%3A3779/datastream/OBJ/view/Comparative_analysis_between_the_diverging_diamond_interchange_and_partial_cloverleaf_interchange_using_microsimulation_modeling.pdf

The Parclo B4 had the shortest delay time and the fewest number of stops for all scenarios tested.  In regards to max queue length, the Parclo B4 had the shortest queues in low volume scenarios but the highest queues in high volume scenarios.  The explanation for the  max queue length of the Parclo B4 is explained on page 92:

The maximum queue length on a ParClo B4 occurs at the off-ramps, where cars are trying to enter the crossroad, but are unable to do so since the crossroad is very congested and the vehicles wanting to go through the intersection do not have to stop since they have green throughout the whole cycle length.  For the High volume flows, the DDI designs had a much better performance than the ParClo designs.The ParClo A4 had double the queue length than the DDIs, and the ParClo B4 had almost triple the maximum queue lengths at some instances compared to the DDIs.

It's incredibly important to note that the Parclo B4 being analyzed in this study had free-flowing loop ramps.  If the off ramps were signalized the max queue of the Parclo B4 would likely be comparable to the other interchanges being analyzed.

The Parclo B4 outperformed the DDI and Parclo A4 in every MOE tested (besides max queue length at high volumes, which can be easily addressed with signalized off-ramps). 

[vtk]

Made possible by the fact that Detroit's arterials are nearly all mile-grid roads with Michigan Lefts.  How much of Boston or Atlanta would have to be demolished to make room for such a neat road system that is so agreeable to signal progression?

If signal progression is already bad, like in the case of Boston or Atlanta, wouldn't that be a good reason to choose interchange designs that improve upon progression?  If someone is drowning would you try to hold their head under the water?

No, it doesn't work like that, and you missed my point. Unless you completely reconfigure the arterial network on a large scale in cities like that, the best signal progression that's even possible is still hardly recognizable as any progression at all. Sure, a choice between interchange designs may have an impact on signal progression, but we're probably talking about two nearly indistinguishable shades of crap, compared to Metro Detroit.

If someone's drowning, do you fuss over whether it's Lake Erie water or Evian that they are drowning in?

What are you arguing here?  I would prefer to drive on a Boston corridor where 90% of the signals have crappy progression as opposed to a Boston corridor where 100% of the signals have crappy progression.  The fact is any city could benefit from interchange designs that improve upon progression.

If a corridor doesn't already have good two-way signal progression, that's probably because the geometric configuration of the arterials doesn't support it. An "interchange design that improves upon progression" at one location along such a corridor isn't going to make any practical difference.  You seem to think there's this additive effect, that every intersection independently contributes to the value of signal progression, but that's nonsense. There's no appreciable value until you have a run of a few arterial—arterial intersections, and then each additional intersection within the run adds more and more value. An interchange that doesn't ruin progression is no better than other interchanges if there's no progression to speak of. It doesn't create progression by itself.

[johndoe]

If the off ramps were signalized the max queue of the Parclo B4 would likely be comparable to the other interchanges being analyzed.

The Parclo B4 outperformed the DDI and Parclo A4 in every MOE tested (besides max queue length at high volumes, which can be easily addressed with signalized off-ramps).

Signalization of the off ramp would disrupt your through coordination, which seems to be your whole argument...

[tradephoric]

Signalization of the off ramp would disrupt your through coordination, which seems to be your whole argument...

Signalizing the off-ramps at a Parclo B4 wouldn't disrupt signal progression as the off-ramp signals would only stop one-direction of travel.  Ever hear of a Restricted Crossing U-Turn intersection (RCUT)?  A Parclo B4 is very similar to an RCUT operationally.

RCUT in Troy Michigan:


Parclo B4 in Dallas, Texas:


One disadvantage of the RCUT is that traffic along the minor road wishing to turn left are not permitted to do so.  They must first make a right at the major road and then make a U-turn.   A Parclo B4 doesn't have this disadvantage as drivers coming off the freeway decide which off-ramp to take so that they are traveling in their desired direction.

In regards to safety, RCUTs have a proven safety record.  Researchers from North Carolina State University have found that RCUTs reduce total accidents by 46% and reduce personal injury accidents by 63% when compared to conventional intersection designs.  Parclo B4's should see similar reductions in injury accidents since the designs are near mirror images. 

[tradephoric]

If a corridor doesn't already have good two-way signal progression, that's probably because the geometric configuration of the arterials doesn't support it. An "interchange design that improves upon progression" at one location along such a corridor isn't going to make any practical difference.  You seem to think there's this additive effect, that every intersection independently contributes to the value of signal progression, but that's nonsense. There's no appreciable value until you have a run of a few arterial—arterial intersections, and then each additional intersection within the run adds more and more value. An interchange that doesn't ruin progression is no better than other interchanges if there's no progression to speak of. It doesn't create progression by itself.

I think people have become apathetic to bad signal progression because it's so common.  Some of the comments on this thread seem to be defending poor signal progression.  If drivers get stopped at every other red light along a corridor, what's the big deal if they get stopped at a SPUI? 

For Detroit drivers who routinely cruise long stretches without getting stuck at red lights, a progression killing interchange stands out like a sore thumb.  It's annoying to cruise 10-miles down a corridor only to get stopped at a progression killing Parclo A4 interchange.  The Parclo A4's are the most annoying to get stopped at since they have similar footprint/costs as a Parclo B4.  If you are going to have a large footprint Parclo, pick the Parclo that maintains good signal progression!  IMO, the Parclo B4 should be the preferred interchange design as opposed to the Parclo A4.

[vtk]

If you are going to have a large footprint Parclo, pick the Parclo that maintains good signal progression!  IMO, the Parclo B4 should be the preferred interchange design as opposed to the Parclo A4.

In Detroit. Almost everywhere else, your argument is weak at best. And I still assert an A4, SPUI, or DDI can potentially be engineered and operated so as not to completely kill progression in a Detroitoid arterial grid.

Tradephoric, can you show me what a signal that only stops one direction of traffic looks like in one of your time/distance diagrams? I want to try to get the visual language right.

[vtk]

Some of the comments on this thread seem to be defending poor signal progression.  If drivers get stopped at every other red light along a corridor, what's the big deal if they get stopped at a SPUI?

That's not defending bad progression, but accepting it. Because if you don't have a neat grid or ROW wide enough for median U turns, good progression is basically impossible anyway. Due to left turn phases and clearance intervals, at each signal with another arterial you have less than a 50% chance of hitting the light on green, and a SPUI will be just another signal with the same chances. But a parclo B4 would only be slightly better on a corridor like this, with maybe better than 50% green probability (depending on the volume of traffic entering the freeway) and likely poor coordination with other signals if loop detectors are used. And if the exit ramps are also signalized to prevent weaving, then each direction of the surface corridor has two signals to deal with in the interchange, and the probability of getting through both of them on green may drop below 50%, so even if the metric is signal progression, the B4 is now comparable to a SPUI.  And there are other things to consider while choosing an interchange design.

[tradephoric]

Tradephoric, can you show me what a signal that only stops one direction of traffic looks like in one of your time/distance diagrams? I want to try to get the visual language right.

Here is an example of a Parclo B4 interchange near Miami, Florida with a corresponding time-distance diagram.


green circle icons = traffic signals that stop only one direction of travel
red circle icons = traffic signals that stop both directions of travel

[johndoe]

Trade, I don't think any of us are saying signal timing isn't important.  But at many service interchanges the through movement isn't the highest volume.  Traffic signals are always going to delay someone...but if the through movement is half the volume of those turning left onto the freeway, why give the through movement priority at the expense of the majority of drivers?

The other thing we need to remember is the arrivals from the off-ramp. What if (assuming north is up in your picture) there are twice as many people going sb to eb as eb to nb?  The green band at Dykes and 148th will be different.  What if there is a huge development on one side of the interchange?   Site conditions are going to vary, so it's not fair to say that just because the through movement doesn't get perfect "waves" the coordination isn't optimized.