Traffic Signals: green wave

[mrsman]

http://chi.streetsblog.org/2015/06/19/surfing-the-green-wave-cdot-pilots-bike-friendly-signal-timing-on-wells/

Here's another clever new idea from the Chicago Department of Transportation. This week, they re-timed the stoplights along Wells Street between Huron Street and Wacker Drive in River North, so that southbound bicyclists who maintain a 12 mph pace get an unbroken series of greens.

Previously, the stoplights were timed so that a cyclist pedaling at 12 mph starting from a green at Huron would hit a red light at almost every intersection. "I think there was a lot of frustration with the number of lights you'd hit and have to wait at, and hopefully that's changed,"  Amsden said. Note that the Wacker Drive stoplight is not part of the wave, so you'll tend to get a red when you arrive there after a series of greens.

The new timing also benefits drivers, Amsden said. If they travel at about 25 mph, they'll also hit every green. That's an incentive for motorists to avoid speeding, and it should also reduce red light running by drivers, making the street safer for everybody. Between 2009 and 2013, there were 35 crashes on this stretch that involved red light running. Amsden added that the new signal timing doesn't hurt traffic flow on major cross streets and, in some cases, it improves it.

My question for the traffic engineers on the board is how do you design signal timing with a green wave at both 12 MPH (bikes) and 25 MPH (cars)?  Note that Wells Street is one-way southbound for this whole stretch.

[Brandon]

CDOT's full of crap here.  The limit on Wells is 30 mph, not 25 mph.  I also don't think you can get bicyclists to accurately know they're going 12 mph as most bicycles lack speedometers.

[jakeroot]

CDOT's full of crap here.  The limit on Wells is 30 mph, not 25 mph.  I also don't think you can get bicyclists to accurately know they're going 12 mph as most bicycles lack speedometers.

Even if the facts are total horseshit, as long as cyclists have an easier time, that's all that matters to most urban DOTs these days.

[tradephoric]

Portland, Oregon time the downtown lights for 12.5 MPH.  This is what it looks like in a simulation model:

https://www.youtube.com/watch?v=7jGWdCknurM

To maintain good signal progression, you gotta maintain 12.5 mph.  A bicyclists who can only average 8 mph will eventually get cut off at the tail end of the wave.  A driver going 25 mph will catch the front of the green wave and eventually hit a red light.

[SidS1045]

ISTR that when NYC's speed limit was 30mph city-wide, the signals on the one-way avenues in Manhattan were set so that a constant speed of 28mph would get you a green at every intersection.

As for the OP, I'm not a traffic engineer, but these days it seems to me that as long as all the signal controllers "talk" to a central computer it shouldn't be too difficult to program a "green wave."

[noelbotevera]

Green waves are great for Atlantic Avenue in downtown Atlantic City, New Jersey. Between the Ventor City line to I'd say the Trump Taj Mahal (with two breaks at Albany Blvd and Missouri Avenue) could use this. The signals are barely concurrent with each other and I do get a tiny bit mad whenever we always have to stop after almost every green. Atlantic Avenue is essentially the Boardwalk's alt route and bypass route, so when I'd like to head towards the Aquarium, I'd want to use Atlantic Avenue instead of using the Boardwalk's north end (which is boring between the Steel Pier and the north end. It's just casino).

[1995hoo]

The one I find odd is on the twin one-way portions of US-1 through the City of Alexandria in Virginia. The speed limit is 25. If you obey that, you will not be able to catch the green wave all the way through, but if you go 30, you'll be in right about the sweet spot if you hit that area at the right time near the start of the green wave.

It just seems peculiar to me to set a green wave such that you have to exceed the speed limit to catch it.

[tradephoric]

Here's a really long green wave along 1st Avenue in NYC.  If you travel 27 MPH you can make it through 125 blocks.  In theory it's not hard to achieve a green wave but in practice to have this many signals stay in step is pretty incredible (having 125 consecutive traffic signals operating how they are suppose to be operating is no small feat).  NYCDOT is doing something right.

https://www.youtube.com/watch?v=Kf720nj0WMk

[roadfro]

My question for the traffic engineers on the board is how do you design signal timing with a green wave at both 12 MPH (bikes) and 25 MPH (cars)?  Note that Wells Street is one-way southbound for this whole stretch.

I'm no traffic engineer, but I did take several grad level traffic engineering courses, so I'll take a stab...

It helps to have a one way street involved, so there aren't any turning phases to worry about. And have that street be in a downtown grid network where signals are evenly spaced is another bonus (but not absolutely necessary).

You need to use a pretty short cycle length to get this right. The main trick here is finding a timing speed such that autos can proceed at one speed and bicyclists go at another and still catch the wave. Note that the wave's bike speed is 12 mph, roughly half of the auto speed. By doing this, a bike will cover the same distance as a car in twice the time.

Assume free-flowing traffic conditions with an evenly-spaced street grid. A car and bike are at the stop line at Street A. When the light turns green, both take off. The car is going to get to Street B as it changes to green. The bike will take twice as long, so it actually arrives at Street B during the second green.

Portland, Oregon time the downtown lights for 12.5 MPH.  This is what it looks like in a simulation model:

<snipped video>

To maintain good signal progression, you gotta maintain 12.5 mph.  A bicyclists who can only average 8 mph will eventually get cut off at the tail end of the wave.  A driver going 25 mph will catch the front of the green wave and eventually hit a red light.

One note there is that the downtown speed limit is 20 or 25, IIRC. The progression wave was deliberately set for slower traffic speeds.

Another thing I noticed with Portland's downtown grid: When walking in the opposite direction of the traffic flow on a one-way street, the progression wave seemed to work in my favor at the pedestrian signals as well.

[roadman65]

Can somebody teach the green way to Orange County, FL Traffic engineering to do this on main drags like Orange Blossom Trail  and John Young Parkway.  I know if you confront the brass on that one, they will say "Impossible."  Their excuse would be the varying left turn signal times that would prevent it.  However, NJDOT did that at Milton Avenue in Rahway, NJ for US 1 & 9.  The third intersecting road used to create a third phase, and sometimes there would be two if there was no traffic on Paterson Street.  Thus the Paterson Street green would just be taken out of the green time for US 1 & 9 and if Paterson had no vehicles call for a signal, then US 1 & 9 would benefit more by having a bonus green time.

All they have to do is use the extended times needed for side roads and left turns to be taken out of the through movements green time.  This will not make it like NYC, as NYC is one way streets and no protected lefts and no major boulevards intersecting, however you might catch somewhat of wave, which is most welcome in Orlando these days.

[DeaconG]

Can somebody teach the green way to Orange County, FL Traffic engineering to do this on main drags like Orange Blossom Trail  and John Young Parkway.  I know if you confront the brass on that one, they will say "Impossible."  Their excuse would be the varying left turn signal times that would prevent it.  However, NJDOT did that at Milton Avenue in Rahway, NJ for US 1 & 9.  The third intersecting road used to create a third phase, and sometimes there would be two if there was no traffic on Paterson Street.  Thus the Paterson Street green would just be taken out of the green time for US 1 & 9 and if Paterson had no vehicles call for a signal, then US 1 & 9 would benefit more by having a bonus green time.

All they have to do is use the extended times needed for side roads and left turns to be taken out of the through movements green time.  This will not make it like NYC, as NYC is one way streets and no protected lefts and no major boulevards intersecting, however you might catch somewhat of wave, which is most welcome in Orlando these days.

Throw in Semoran Blvd, Colonial and Alafaya Trail and I'm with you!

[roadman65]

Can somebody teach the green way to Orange County, FL Traffic engineering to do this on main drags like Orange Blossom Trail  and John Young Parkway.  I know if you confront the brass on that one, they will say "Impossible."  Their excuse would be the varying left turn signal times that would prevent it.  However, NJDOT did that at Milton Avenue in Rahway, NJ for US 1 & 9.  The third intersecting road used to create a third phase, and sometimes there would be two if there was no traffic on Paterson Street.  Thus the Paterson Street green would just be taken out of the green time for US 1 & 9 and if Paterson had no vehicles call for a signal, then US 1 & 9 would benefit more by having a bonus green time.

All they have to do is use the extended times needed for side roads and left turns to be taken out of the through movements green time.  This will not make it like NYC, as NYC is one way streets and no protected lefts and no major boulevards intersecting, however you might catch somewhat of wave, which is most welcome in Orlando these days.

Throw in Semoran Blvd, Colonial and Alafaya Trail and I'm with you!

Oh yes.  I thought of them too, but did not want to get lengthy.

[tradephoric]

Can somebody teach the green way to Orange County, FL Traffic engineering to do this on main drags like Orange Blossom Trail  and John Young Parkway.  I know if you confront the brass on that one, they will say "Impossible."  Their excuse would be the varying left turn signal times that would prevent it.  However, NJDOT did that at Milton Avenue in Rahway, NJ for US 1 & 9.  The third intersecting road used to create a third phase, and sometimes there would be two if there was no traffic on Paterson Street.  Thus the Paterson Street green would just be taken out of the green time for US 1 & 9 and if Paterson had no vehicles call for a signal, then US 1 & 9 would benefit more by having a bonus green time.

All they have to do is use the extended times needed for side roads and left turns to be taken out of the through movements green time.  This will not make it like NYC, as NYC is one way streets and no protected lefts and no major boulevards intersecting, however you might catch somewhat of wave, which is most welcome in Orlando these days.

Throw in Semoran Blvd, Colonial and Alafaya Trail and I'm with you!

Oh yes.  I thought of them too, but did not want to get lengthy.

There are over 70 intersections in the Orlando area that have pedestrian crosswalks of 150 feet or more.  To fit the pedestrian intervals, many of the traffic signals in Orlando need to run 180 second cycle lengths (to adhere to MUTCD standards).  Short of redesigning the intersections with shorter crosswalks, 180 second cycles is baked into the cake.

Long cycle lengths are problematic since they require signals to be spaced farther apart to achieve dual signal progression.  The critical resonant cycle length is a cycle that can achieve good dual progression and is equal to 2 x travel time (where travel time = distance / speed).  Assuming a 45 mph corridor, we can plug in the numbers to see how far the signals need to be spaced to maintain dual signal progression:

Critical resonant cycle = 2 * distance / speed

180 = 2 * distance / [(45mph* 3600s/h) / 5280 ft/mi)]

So distance = 5,940 feet

Basically, if the signals in Orlando are spaced closer than 5,940 feet apart, perfect dual signal progression is impossible (assuming theoretically 50/50 splits).  Here's a time-distance to illustrate the good progression you get when the signals are spaced 5,940 feet apart.  OTOH, if the signals are spaced 3,000 feet apart, you end up with the worst possible dual progression possible. 





So how far are the major signals spaced apart along Orange Blossom Trail?  Traffic engineers can play with the splits and lead/lag phasing to improve signal progression somewhat, but the farther you are from theoretically perfect dual progression the worst progression will be.

[jakeroot]

Short of redesigning the intersections with shorter crosswalks, 180 second cycles is baked into the cake.

Shouldn't 180 seconds only be required if pedestrians are present? Worse case scenario, you could build pedestrian tunnels beneath the intersection.

[iBallasticwolf2]

Short of redesigning the intersections with shorter crosswalks, 180 second cycles is baked into the cake.

Shouldn't 180 seconds only be required if pedestrians are present? Worse case scenario, you could build pedestrian tunnels beneath the intersection.

Expensive but safe and convienent for pedestrians.

[tradephoric]

Short of redesigning the intersections with shorter crosswalks, 180 second cycles is baked into the cake.

Shouldn't 180 seconds only be required if pedestrians are present? Worse case scenario, you could build pedestrian tunnels beneath the intersection.

Expensive but safe and convienent for pedestrians.

Take pedestrians out of the equation.  A typical Orlando intersection is 4-phases with protected left turns.  With this type of setup, you would still need to run 120 second cycles to effectively service the vehicle demands (as shorter cycles lead to higher start up lost time).  Plug in 120 second into the critical resonant cycle formula and you end up with an ideal signal spacing of 3,960 feet.  Another problem Orlando has is the signal spacing aren't consistent to begin with, making selecting an optimum cycle length along a corridor difficult since it's always a moving target.

A possible solution for Orlando is constructing MUTs (Michigan Lefts) and RCUTs (superstreets) along the major corridors.  These designs split up pedestrian crossings into two shorter crossings, so Orlando would no longer be hijacked by 150+ feet pedestrian crossings.  MUT and RCUT designs also lead to simple 2-phase signals, eliminating all left-turn phases from the corridor.  This increases the capacity and effective green bands at each traffic signal.   A combination of MUT and RCUT designs can be extremely effective at keeping traffic moving. 

[jakeroot]

A possible solution for Orlando is constructing MUTs (Michigan Lefts) and RCUTs (superstreets) along the major corridors.  These designs split up pedestrian crossings into two shorter crossings, so Orlando would no longer be hijacked by 150+ feet pedestrian crossings.  MUT and RCUT designs also lead to simple 2-phase signals, eliminating all left-turn phases from the corridor.  This increases the capacity and effective green bands at each traffic signal.   A combination of MUT and RCUT designs can be extremely effective at keeping traffic moving.

How familiar are you with Europe? I wonder how they approach issues like this. Do they also use MUTs and RCUTs? They certainly have less grids than we do, but they also have at-grade expressways, much like the roads criss-crossing suburban Florida.

[DaBigE]

Short of redesigning the intersections with shorter crosswalks, 180 second cycles is baked into the cake.

Shouldn't 180 seconds only be required if pedestrians are present? Worse case scenario, you could build pedestrian tunnels beneath the intersection.

Expensive but safe and convienent for pedestrians.

Safe, as long as they're lit properly and people feel safe using them. All it takes is one crime to happen...

[Brian556]

Quote from roadman 65:

Can somebody teach the green way to Orange County, FL Traffic engineering to do this on main drags like Orange Blossom Trail  and John Young Parkway.  I know if you confront the brass on that one, they will say "Impossible."  Their excuse would be the varying left turn signal times that would prevent it.  However, NJDOT did that at Milton Avenue in Rahway, NJ for US 1 & 9.  The third intersecting road used to create a third phase, and sometimes there would be two if there was no traffic on Paterson Street.  Thus the Paterson Street green would just be taken out of the green time for US 1 & 9 and if Paterson had no vehicles call for a signal, then US 1 & 9 would benefit more by having a bonus green time.

All they have to do is use the extended times needed for side roads and left turns to be taken out of the through movements green time.  This will not make it like NYC, as NYC is one way streets and no protected lefts and no major boulevards intersecting, however you might catch somewhat of wave, which is most welcome in Orlando these days.

Never mind the whole green wave thing, Orange County just doesn't know how to time signals at all. At 535 and 536, the green light was only letting 10 cars thru at a time, and traffic was backed up about a mile. Yes, the intersection is overloaded, but good signal timing could greatly improve things.
Another huge part of the problem in the Orlando area is the lack of a proper arterial street grid. See the northern suburbs of Dallas, TX for a good example.

This topic brings to mind another thought. Are we doing it all wrong by having these huge divided streets in the suburbs that require too many signal phases? Would it be better to use one way pairs, and put businesses in between the one way streets to make them easily accessible from both streets?
This would also make it easier to cross or turn left at unsignalized locations, and would greatly reduce accidents.
The most common cause of accidents on these arterials is drivers attempting to cross or turn left at unsignalized locations. In my opinion, these unsafe movements should not be possible. Florida does way better than Texas on this.

[tradephoric]

This topic brings to mind another thought. Are we doing it all wrong by having these huge divided streets in the suburbs that require too many signal phases? Would it be better to use one way pairs, and put businesses in between the one way streets to make them easily accessible from both streets?
This would also make it easier to cross or turn left at unsignalized locations, and would greatly reduce accidents.

How about a Town Center Intersection (TCI)?  A simple 2-phase design with direct left turns.  It's essentially a two-way street that widens out to a one-way pair as you approach a major intersection.
https://www.youtube.com/watch?v=rCQn89wItTY

Route 22 near Union, NJ is similar to what you are describing as well. It's basically an extra wide RCUT with commercial development in the median.

https://www.google.com/maps/@40.68923,-74.30305,290m/data=!3m1!1e3

[Mdcastle]

What's dual signal progression?
What's happening in the charts? Is this traffic each direction on one street? On the second chart does traffic on the street one direction hit two greens and the other direction hit a red?

[tradephoric]

What's dual signal progression?
What's happening in the charts? Is this traffic each direction on one street? On the second chart does traffic on the street one direction hit two greens and the other direction hit a red?

It's difficult to time signals along a two-way street.  Timing the signals so drivers hit a wave of green lights in one direction may lead to drivers in the opposing direction hitting every red light (ie. the second chart).  Dual signal progression is when drivers are able to hit a wave of green lights in BOTH directions of travel (ie. the first chart). 

Here's are some factors that kill good dual progression along a corridor:
1.  Irregularly spaced traffic signals.
2.  Traffic signals spaced too closely together.

It's important to note that sometimes "green waves" aren't the best type of progression.  In heavy traffic or when signals are spaced very close together, simultaneous greens can help clear out traffic queues between signals.  That said, it's always nice to be able to get a green wave for both directions.

[tradephoric]

It helps to have a one way street involved, so there aren't any turning phases to worry about. And have that street be in a downtown grid network where signals are evenly spaced is another bonus (but not absolutely necessary).

You need to use a pretty short cycle length to get this right. The main trick here is finding a timing speed such that autos can proceed at one speed and bicyclists go at another and still catch the wave. Note that the wave's bike speed is 12 mph, roughly half of the auto speed. By doing this, a bike will cover the same distance as a car in twice the time.

Assume free-flowing traffic conditions with an evenly-spaced street grid. A car and bike are at the stop line at Street A. When the light turns green, both take off. The car is going to get to Street B as it changes to green. The bike will take twice as long, so it actually arrives at Street B during the second green.

Assuming a 60 second cycle, the signals would have to be spaced 2,200 feet apart before the bicyclist and motorists could achieve a perfect green wave at 12.5 mph and 25 mph respectively.  However, if the bicyclist only averages 10 mph over those 2,200 feet, they would experience the worst signal progression possible.

Here's another video of Wells Street from a bicyclist perspective.  The bicyclist enjoys a steady green wave at 12 mph while the vehicles experience stop-go-stop-go progression at 25 mph.  The bicyclists pretty much keep up with the platoon of vehicles.  This makes sense since the street is timed for 12 mph.  Simply put, the claim the guy from CDOT made is wrong.

https://www.youtube.com/watch?v=h5D2S7tP9Y8

[roadfro]

It helps to have a one way street involved, so there aren't any turning phases to worry about. And have that street be in a downtown grid network where signals are evenly spaced is another bonus (but not absolutely necessary).

You need to use a pretty short cycle length to get this right. The main trick here is finding a timing speed such that autos can proceed at one speed and bicyclists go at another and still catch the wave. Note that the wave's bike speed is 12 mph, roughly half of the auto speed. By doing this, a bike will cover the same distance as a car in twice the time.

Assume free-flowing traffic conditions with an evenly-spaced street grid. A car and bike are at the stop line at Street A. When the light turns green, both take off. The car is going to get to Street B as it changes to green. The bike will take twice as long, so it actually arrives at Street B during the second green.

Assuming a 60 second cycle, the signals would have to be spaced 2,200 feet apart before the bicyclist and motorists could achieve a perfect green wave at 12.5 mph and 25 mph respectively.  However, if the bicyclist only averages 10 mph over those 2,200 feet, they would experience the worst signal progression possible.

Here's another video of Wells Street from a bicyclist perspective.  The bicyclist enjoys a steady green wave at 12 mph while the vehicles experience stop-go-stop-go progression at 25 mph.  The bicyclists pretty much keep up with the platoon of vehicles.  This makes sense since the street is timed for 12 mph.  Simply put, the claim the guy from CDOT made is wrong.

I guess I didn't take distance into account in my reply. Not being familiar with the Chicago grid system, I wonder how far apart are their streets are. Is there a reasonable cycle length that could achieve the green wave for both bikes and cars at these speeds?

[mrsman]

It helps to have a one way street involved, so there aren't any turning phases to worry about. And have that street be in a downtown grid network where signals are evenly spaced is another bonus (but not absolutely necessary).

You need to use a pretty short cycle length to get this right. The main trick here is finding a timing speed such that autos can proceed at one speed and bicyclists go at another and still catch the wave. Note that the wave's bike speed is 12 mph, roughly half of the auto speed. By doing this, a bike will cover the same distance as a car in twice the time.

Assume free-flowing traffic conditions with an evenly-spaced street grid. A car and bike are at the stop line at Street A. When the light turns green, both take off. The car is going to get to Street B as it changes to green. The bike will take twice as long, so it actually arrives at Street B during the second green.

Assuming a 60 second cycle, the signals would have to be spaced 2,200 feet apart before the bicyclist and motorists could achieve a perfect green wave at 12.5 mph and 25 mph respectively.  However, if the bicyclist only averages 10 mph over those 2,200 feet, they would experience the worst signal progression possible.

Here's another video of Wells Street from a bicyclist perspective.  The bicyclist enjoys a steady green wave at 12 mph while the vehicles experience stop-go-stop-go progression at 25 mph.  The bicyclists pretty much keep up with the platoon of vehicles.  This makes sense since the street is timed for 12 mph.  Simply put, the claim the guy from CDOT made is wrong.

I guess I didn't take distance into account in my reply. Not being familiar with the Chicago grid system, I wonder how far apart are their streets are. Is there a reasonable cycle length that could achieve the green wave for both bikes and cars at these speeds?

This is taking place in the River North neighborhood of Chicago just north of Chicago's Loop (Downtown).  There is a light at every block and the blocks in this section are about  1/16 of a mile, or 330 feet.

I agree with tradephoric, it seems that they simply timed the lights for bikes and not really accounted for cars.  So in answer to my original question, it seems that it is not possible to time streets for two separate speeds of progression at the same time, when the signals are so closely spaced together.