Intersection help.

[Amtrakprod]

The intersection with Lexington st and Trapelo Road has been a traffic nightmare for over ten years. Does anyone have any ideas about how to fix it. Here's the Google Map link: https://www.google.com/maps/@42.410668,-71.233806,3a,75y,23.52h,91.19t/data=!3m6!1e1!3m4!1smWageZJxxIpClT_hbqMBXw!2e0!7i13312!8i6656
The cycle of the light goes like this:
Green left arrows on Lexington st, then they both go red. Lexington st gets a green except left arrows. Then Green left arrows on Trapelo Road, after they go red. Trapelo Road then gets a green except for left turns. Then all crosswalks change.
Anyways, would like to see your answers with timing or geometry, whatever y'all come up with, I'll enjoy to read.




iPhone

[hotdogPi]

Make the Mobil gas station entrance/exit a slip ramp, instead of having a combined straight/right turn from the north.

[Amtrakprod]

Make the Mobil gas station entrance/exit a slip ramp, instead of having a combined straight/right turn from the north.

Might help, but I don't think that will really work super well, I think some timing has to be changed.


iPhone

[jakeroot]

I'm working on a redesign of the intersection right now. I will upload it when I'm done. It includes new signals and lane assignment modifications.

Which are the busier movements? Just to make sure I'm doing this right.

[Amtrakprod]

Which are the busier movements? Just to make sure I'm doing this right.

East Trapelo Road straight/left and North Lexington st Straight/Right.



iPhone

[jeffandnicole]

It's probably just better signal timings.  Geometry seems OK; it's a perfect + intersection.  Maybe lengthen the left turn slots a bit and provide right turn slots.

It doesn't appear it needs a full scale reconstruction - signal timings just need to be updated for today's traffic - something that probably hasn't been adjusted in over a decade.

[Amtrakprod]

It's probably just better signal-signal timings just need to be updated for today's traffic - something that probably hasn't been adjusted in over a decade.

Do you have a suggested timing?


iPhone

[jeffandnicole]

It's probably just better signal-signal timings just need to be updated for today's traffic - something that probably hasn't been adjusted in over a decade.

Do you have a suggested timing?


iPhone

First, we need to know what the timings are now.  We also need to know the traffic volumes, especially at rush hours.  We also need to know what other conditions exist in the area, how far away the next traffic lights are, etc.

Generally speaking, longer is better, but you have to factor in queue lengths and other stuff.  You could have 5 minute cycles, but if the left turn lane backs up into the thru lane, the benefit of a long green is killed because traffic bottlenecks at the beginning of the left turn lane.  Without knowing any of the above facts, 2 - 3 minute cycles tend to work pretty well.

[ET21]

I have a feeling it's definitely signal timings. Overall I like the intersection and the lane configs. I hate it when they fuck around with the light timings because it can screw up your traffic badly (Cicero Ave and Lake Shore Drive in Chicago come to mind immediately)

[tradephoric]

First, I would widen Lexington Street to a 5-lane road with dedicated center-left turn lanes.  This would increase left-turn queue space and help prevent left turners from blocking through traffic on Lexington Street. 

Second, I would adjust the signal timings.  You describe that the lefts "both go red"  suggesting that the lefts terminate together.  Each left turn movement should be allowed to gap out early, and any unused time given to the opposing through phase.   If the left-turns always terminate together, that's definitely not ideal for maximizing capacity at the intersection.

Generally speaking, longer is better, but you have to factor in queue lengths and other stuff.  You could have 5 minute cycles, but if the left turn lane backs up into the thru lane, the benefit of a long green is killed because traffic bottlenecks at the beginning of the left turn lane.  Without knowing any of the above facts, 2 - 3 minute cycles tend to work pretty well.

I agree with J&N (shocking I know!) that longer cycles generally increase intersection capacity. The problem with short cycles is that there is too much wasted time to fit the clearance intervals.  If a 4-phase signal is running a short 60 second cycle, nearly 40% of the cycle would be dedicated to running clearance intervals assuming 6 second clearances each phase (6 second yellow/red/start up loss X 4 phases = 24 seconds / 60 second cycle = 40%).  Now there is a problem with long cycles lengths too though.  If the cycle length is too long, the saturation flow rate begins to drop as you get deeper into the green phase.  Research has found that about 40-50 second phase green times maximizes capacity so for a simple 2-phase signal with no left turn phases, 80-100 second cycles would be close to ideal.  But once you got to service left-turn phases, 120 second cycles is really about the minimum cycle you would want to run during peak periods.  Can you let us know what cycle the signal is actually running during the heavy rush?  Just take a stop watch and see how long it takes for the signal to cycle through all the phases.  The cycle length will very likely be a multiple of 10.

[hotdogPi]

I'm not sure if this would work, but is there any way to time this signal and the Smith St./Trapelo Rd. intersection so that someone going through both (in either direction) will always get a green light for the second one?

[tradephoric]

I'm not sure if this would work, but is there any way to time this signal and the Smith St./Trapelo Rd. intersection so that someone going through both (in either direction) will always get a green light for the second one?

I wouldn't be too worried about coordination with the signal at Smith St/Trapelo Road.  That signal is over a mile away and Trapelo Road is a 2-lane road with a 35 mph speed limit.  If you get stuck behind a guy going exactly the speed limit, it would take 102 seconds to travel that mile.  The next day you are cruising 43 mph and it only takes 83 seconds to travel that mile.  That's about a 20 second difference... which day should you coordinate the traffic signals for?  It's likely that the through time for Trapelo Road is only about 20 seconds long as it is (seems the left turn movements onto Lexington would be heavier than the through).  Coordination, IMO, is not very important between those two signals. 

[hotdogPi]

I'm not sure if this would work, but is there any way to time this signal and the Smith St./Trapelo Rd. intersection so that someone going through both (in either direction) will always get a green light for the second one?

I wouldn't be too worried about coordination with the signal at Smith St/Trapelo Road.  That signal is over a mile away and Trapelo Road is a 2-lane road with a 35 mph speed limit.  If you get stuck behind a guy going exactly the speed limit, it would take 102 seconds to travel that mile.  The next day you are cruising 43 mph and it only takes 83 seconds to travel that mile.  That's about a 20 second difference... which day should you coordinate the traffic signals for?  It's likely that the through time for Trapelo Road is only about 20 seconds long as it is (seems the left turn movements onto Lexington would be heavier than the through).  Coordination, IMO, is not very important between those two signals.

What about coordinating it with the signal to the north (Concord Ave/Waltham St.) instead?

[PHLBOS]

I'm not sure if this would work, but is there any way to time this signal and the Smith St./Trapelo Rd. intersection so that someone going through both (in either direction) will always get a green light for the second one?

I wouldn't be too worried about coordination with the signal at Smith St/Trapelo Road.  That signal is over a mile away and Trapelo Road is a 2-lane road with a 35 mph speed limit.  If you get stuck behind a guy going exactly the speed limit, it would take 102 seconds to travel that mile.  The next day you are cruising 43 mph and it only takes 83 seconds to travel that mile.  That's about a 20 second difference... which day should you coordinate the traffic signals for?  It's likely that the through time for Trapelo Road is only about 20 seconds long as it is (seems the left turn movements onto Lexington would be heavier than the through).  Coordination, IMO, is not very important between those two signals.

What about coordinating it with the signal to the north (Concord Ave/Waltham St.) instead?

You are aware that such is over a half-a-mile away and in another town (Lexington)?  Once in Lexington, Waltham St. drops down to a mostly 2-laner and has a lower thickly-settled speed limit of 25 (vs. the state's standard of 30).

[tradephoric]

While Concord Ave is a lot closer (only about a 1/2 mile), Lexington/Waltham still narrows down to two-lane road North of Trapelo Road and I see a 25mph posted speed limit on streetview.  Unless if the signals are extremely close together, worrying about coordination along a 2-lane road with a 25 mph speed limit is not that important.    EDIT:  must have posted this right as PHLBOS responded... which is why i basically just repeated all his same points!

I would worry about keeping coordination with Lexington / Trapelo and the signal about 0.6 miles to the south at Lexington / Bishops Forest Dr.  This would be the most critical signal to keep good coordination IMO. 

[hotdogPi]

I would worry about keeping coordination with Lexington / Trapelo and the signal about 0.6 miles to the south at Lexington / Bishops Forest Dr.  This would be the most critical signal to keep good coordination IMO.

Lexington/Trapelo and Lexington/Bishops Forest Dr. are 4 traffic lights apart. There are 3 in between.

[tradephoric]

I would worry about keeping coordination with Lexington / Trapelo and the signal about 0.6 miles to the south at Lexington / Bishops Forest Dr.  This would be the most critical signal to keep good coordination IMO.

Lexington/Trapelo and Lexington/Bishops Forest Dr. are 4 traffic lights apart. There are 3 in between.

Wow you are right!  I can't believe i missed that many traffic signals looking at the aerial.  I will say the signal at Stearns Hill Road, directly south of Traepelo Road, is so close to the main intersection that you could maintain good progression in both directions.   Similarly, Brookway Road is so close to Bishops Forest Drive that they you could maintain good progression as well.  The signal that will likely stop drivers is the one at Lionel Ave, which appears to service some apartment complexes.  So the critical signals to coordinate along Lexington would be Trapelo, Lionel, and Bishops Forest Drive (but i'll tell you right now just based on the signals spacings and assuming a required 120 second cycle, two-way progression is not going to look very good).

[tradephoric]

Here's what different signal progression setups might look like along Lexington Street between Lake St/Bishops Forsest Dr and Trapelo.  You could set up good signal progression for one direction of travel (top time-distance) but then the other direction comes straight to a red at Lionel Ave.  If you set it up for dual progression, then both directions get cut off at the main signals but will likely make it through the minors (bottom time-distance).

[roadfro]

The intersection with Lexington st and Trapelo Road has been a traffic nightmare for over ten years. Does anyone have any ideas about how to fix it. Here's the Google Map link: https://www.google.com/maps/@42.410668,-71.233806,3a,75y,23.52h,91.19t/data=!3m6!1e1!3m4!1smWageZJxxIpClT_hbqMBXw!2e0!7i13312!8i6656
The cycle of the light goes like this:
Green left arrows on Lexington st, then they both go red. Lexington st gets a green except left arrows. Then Green left arrows on Trapelo Road, after they go red. Trapelo Road then gets a green except for left turns. Then all crosswalks change.
Anyways, would like to see your answers with timing or geometry, whatever y'all come up with, I'll enjoy to read.

There's the problem right there... Does this intersection seriously run the pedestrian phases completely separate from vehicle traffic? That doesn't seem necessary, especially if there is not any diagonal crossing going on. The only time I would expect to see a "scramble" phase is in a busy downtown area with really high volumes of pedestrians (like multiple peds every cycle during peak and moderate times), and this intersection doesn't seem to fit the bill.

Change the signal timing to allow pedestrian phases concurrently with the adjacent through green. This will require modification of the signal heads to remove the right green arrows in all directions. You'd also need to install additional pedestrian push buttons, as it currently looks like there's only one button on each corner and you'd need one button for each crosswalk.


Also, it seems like this intersection has protected left turns in all directions. I don't see a reason why FYAs couldn't be installed.

[SectorZ]

The intersection with Lexington st and Trapelo Road has been a traffic nightmare for over ten years. Does anyone have any ideas about how to fix it. Here's the Google Map link: https://www.google.com/maps/@42.410668,-71.233806,3a,75y,23.52h,91.19t/data=!3m6!1e1!3m4!1smWageZJxxIpClT_hbqMBXw!2e0!7i13312!8i6656
The cycle of the light goes like this:
Green left arrows on Lexington st, then they both go red. Lexington st gets a green except left arrows. Then Green left arrows on Trapelo Road, after they go red. Trapelo Road then gets a green except for left turns. Then all crosswalks change.
Anyways, would like to see your answers with timing or geometry, whatever y'all come up with, I'll enjoy to read.

There's the problem right there... Does this intersection seriously run the pedestrian phases completely separate from vehicle traffic? That doesn't seem necessary, especially if there is not any diagonal crossing going on. The only time I would expect to see a "scramble" phase is in a busy downtown area with really high volumes of pedestrians (like multiple peds every cycle during peak and moderate times), and this intersection doesn't seem to fit the bill.

Change the signal timing to allow pedestrian phases concurrently with the adjacent through green. This will require modification of the signal heads to remove the right green arrows in all directions. You'd also need to install additional pedestrian push buttons, as it currently looks like there's only one button on each corner and you'd need one button for each crosswalk.


Also, it seems like this intersection has protected left turns in all directions. I don't see a reason why FYAs couldn't be installed.

I can tell you there are not a lot of pedestrians crossing at that intersection. Further south down Lexington St there are plenty of pedestrians at the signals further south.

Also, Lexington Rd is a complete disaster for non-auto traffic. It is one of the rare roads I outright avoid in rush-hour while cycling.

[Amtrakprod]

Also, it seems like this intersection has protected left turns in all directions. I don't see a reason why FYAs couldn't be installed.

I think that idea would be smart, especially if protected left turns was only during Rush Hour. I'm excited to see what Jake comes up with.

[jakeroot]

Also, it seems like this intersection has protected left turns in all directions. I don't see a reason why FYAs couldn't be installed.

I think that idea would be smart, especially if protected left turns was only during Rush Hour. I'm excited to see what Jake comes up with.

I'm putting the final touches on it right now. It's not wildly different than what already exists, but the changes are good IMO. The 3D rendering takes a while.

[Amtrakprod]

I'm putting the final touches on it right now. It's not wildly different than what already exists, but the changes are good IMO. The 3D rendering takes a while.

Great!



iPhone

[jakeroot]

Welp, here you go! Took me forever; school kept getting in my way .

I hate protected lefts, so I designed the intersection around flashing yellow arrows. The WB and EB approaches were changed to double lefts with FYAs, and the SB and NB left turns remain single left turns with FYAs. Two left turn yield lanes are very unusual in the US, but can be safe if designed properly (which mine mostly are -- I would have used wider medians in retrospect). The EB approach's right lane was changed to right only, to allow for a constant free NB to EB right turn. Slip lane added on the NW corner near the gas station. EB and WB approaches have right turn (5-section) filter signals to maximize turn capacity. As mentioned, the other two right turns are not signal-controlled.

Note that the left turn lanes are to the left of the median. This improves visibility of oncoming traffic by creating a gap between through and left turns. I used 1-metre wide medians here, but I'd rather use 2-metre wide medians instead.

All lanes are 11 feet wide. No reason for 12 foot lanes here, IMO.

Everything you see, except the signs and the arrows, were made by myself. Sorry for the poor-quality satellite photo...SketchUp needs some work.

Only thing not pictured are pedestrian signal heads and mast lighting. Still struggling with those.

Overhead views...




Up-close view of some signals...



Northbound approach...



Westbound approach...



Southbound approach...



Eastbound approach...

[jakeroot]

I think we've gotten a bit too far off-topic. I'd rather stay focused on my proposal. I don't mind brief tangents, but I don't want this descending into another "America sucks at driving" thread. We have enough of those.

Amtrakprod, I'm interested in seeing your thoughts when you return.