Two Lanes Merging, Only One Gets Its Own Lane

[mcmc]

Can any engineer explain why the center lanes merge arrangement would ever be preferable to having the right- or left-most lane drop?

[paulthemapguy]

Can any engineer explain why the center lanes merge arrangement would ever be preferable to having the right- or left-most lane drop?

If the outer lane is the one that terminates first, then there is no point to having two lanes on the ramp, because the outer lane would have to merge into the other lane entering the expressway.  That would be functionally the same as having two lanes on an entrance ramp merging into one before reaching the expressway's mainline.

Alternately, think about it this way:  when the center lanes merge, as in the situation you mentioned, a vehicle in the right lane of the mainline expressway can get out of the way and scoot one lane to the left.  In the opposite case, where the outer lane of cars terminates right away, who has to get out of the way of the cars jumping ship from the ending lane?  It's the people in the inner lane of the entrance ramp, who aren't sure of where to go yet because they're just now becoming acquainted with mainline traffic.  It would be better if the left lane of the entrance ramp went through a merge situation similar to that of a single lane merging onto the highway, then allowing the outer lane's merge onto the mainline to occur at a later time. 

[jeffandnicole]

Can any engineer explain why the center lanes merge arrangement would ever be preferable to having the right- or left-most lane drop?

If the outer lane is the one that terminates first, then there is no point to having two lanes on the ramp, because the outer lane would have to merge into the other lane entering the expressway.  That would be functionally the same as having two lanes on an entrance ramp merging into one before reaching the expressway's mainline.

Alternately, think about it this way:  when the center lanes merge, as in the situation you mentioned, a vehicle in the right lane of the mainline expressway can get out of the way and scoot one lane to the left.  In the opposite case, where the outer lane of cars terminates right away, who has to get out of the way of the cars jumping ship from the ending lane?  It's the people in the inner lane of the entrance ramp, who aren't sure of where to go yet because they're just now becoming acquainted with mainline traffic.  It would be better if the left lane of the entrance ramp went through a merge situation similar to that of a single lane merging onto the highway, then allowing the outer lane's merge onto the mainline to occur at a later time. 

In 99.999999% of the situations, the situation that you say "people aren't sure of which way to go yet" is the situation that occurs.

It's pretty obvious that if you're merging onto a highway, you merge into the lanes that already exist, not the lane that co-existed on the rmap.

[Mapmikey]

Almost got into a wreck at night in Dallas a few years back at this one:  https://goo.gl/maps/9B19p8khU1R2

Regular merge sign, and coming around a sharp curve, and I had never been here before to know to look for somebody in a suicide merge.

This interchange is being torn up so this is not like this right now...

The US 52 to I-26 connector heading onto I-26 east (Exit 208) used to have one of these.  You were risking your life if you used the left lane of the connector ramp.

[J N Winkler]

An inside lane merge (assuming traffic circulating on the right, merging leftward) is a rational choice if traffic in the far-right lane on the freeway mainline is reliably less dense than traffic in the far-right lane on the ramp.  This is not to say that it doesn't cause problems that don't require mitigation.  Tiger tails are one approach that has been extensively tested in Britain and works quite well; I am mildly skeptical of the various warning-sign designs that have been tried in the US.  Ultimately, I think locals and frequent visitors acclimatize to inside lane merges as a regionalism, which is what they largely are in the US.  I also suspect that the jurisdictions that use inside lane merges are more likely to attempt to socialize drivers to expect them by using absent-nose tapers at simple merges.

[mrsman]

An inside lane merge (assuming traffic circulating on the right, merging leftward) is a rational choice if traffic in the far-right lane on the freeway mainline is reliably less dense than traffic in the far-right lane on the ramp.  This is not to say that it doesn't cause problems that don't require mitigation.  Tiger tails are one approach that has been extensively tested in Britain and works quite well; I am mildly skeptical of the various warning-sign designs that have been tried in the US.  Ultimately, I think locals and frequent visitors acclimatize to inside lane merges as a regionalism, which is what they largely are in the US.  I also suspect that the jurisdictions that use inside lane merges are more likely to attempt to socialize drivers to expect them by using absent-nose tapers at simple merges.

The real problem with the inside lane merge in the way that it exists in most places in the US is that there isn't enough time (space) to merge easily.  The examples are so abrupt that it leads to something that is quite dangerous.

The best way to handle this is to allow more time for the merge.  The Australian example upthread was a great example of using tapering to allow for more space.

Another example is the merge of the southbound NJ Turnpike car lanes and truck lanes around Exit 6.  The 3-lane car lanes and the 3-lane truck lanes are like separate highways that merge at this point to form a single 3-lane carriageway.  The left lane of the truck lanes and the right lane of the car lanes merges to form the middle lane of the merged highway, but it does not happen abruptly.  Traffic had about 1,200 feet to make the merge and it happens relatively smoothly.   


https://www.google.com/maps/@40.0792016,-74.7494773,3a,75y,240.16h,78.3t/data=!3m6!1e1!3m4!1s10zGOOZcv_24g0lmV9Soag!2e0!7i13312!8i6656!6m1!1e1

[paulthemapguy]

The real problem with the inside lane merge in the way that it exists in most places in the US is that there isn't enough time (space) to merge easily.  The examples are so abrupt that it leads to something that is quite dangerous.

The best way to handle this is to allow more time for the merge.  The Australian example upthread was a great example of using tapering to allow for more space.

This is absolutely right.  An inside lane merge is necessary, because otherwise you negate the utility of the rightmost lane entirely.  So the best approach is to ensure that the inside lane merge is carried out as safely as possible, giving a nice long taper so that drivers have the utmost time to assess the situation and make their move.  The ramp from NB I-355 to EB I-88 is a nice example:  a nice, long approach, leading up to a nice long inside lane merge as you move forward, allowing plenty of time for drivers to assess and react.

https://www.google.com/maps/@41.809586,-88.047163,3a,75y,87.43h,91.13t/data=!3m7!1e1!3m5!1sznjCGERmYU7nUiOlorh_0A!2e0!6s%2F%2Fgeo2.ggpht.com%2Fcbk%3Fpanoid%3DznjCGERmYU7nUiOlorh_0A%26output%3Dthumbnail%26cb_client%3Dmaps_sv.tactile.gps%26thumb%3D2%26w%3D203%26h%3D100%26yaw%3D328.6588%26pitch%3D0%26thumbfov%3D100!7i13312!8i6656

[johndoe]

An inside lane merge is necessary, because otherwise you negate the utility of the rightmost lane entirely.

I think I would disagree with this, from a capacity standpoint.  Regardless the taper rate of an inside merge (or ability to "see it coming ") the right of way isn't clear.  Which lane is supposed to yield? This lack of clarity probably impacts the capacity more than a lane drop and parallel acceleration lane.

[Brandon]

An inside lane merge is necessary, because otherwise you negate the utility of the rightmost lane entirely.

I think I would disagree with this, from a capacity standpoint.  Regardless the taper rate of an inside merge (or ability to "see it coming ") the right of way isn't clear.  Which lane is supposed to yield? This lack of clarity probably impacts the capacity more than a lane drop and parallel acceleration lane.

The vehicles merging onto the freeway from the lane that merges, should either yield or speed up while finding their gap in traffic.  Those in the lane that gets its own new lane, can do as they please.

[Ned Weasel]

I think I would disagree with this, from a capacity standpoint.  Regardless the taper rate of an inside merge (or ability to "see it coming ") the right of way isn't clear.  Which lane is supposed to yield? This lack of clarity probably impacts the capacity more than a lane drop and parallel acceleration lane.

Drivers shouldn't assume they have the right of way at any merge.  If one is explicitly told to yield, then, yes, one should yield.  But in a merge where neither side has a yield sign, drivers should have enough sense to act courteously and cautiously when they see that two vehicles are trying to occupy the same space ahead (slow down a little if necessary, but don't stop if you can avoid it, and don't play chicken).  At any rate, this would rarely be an issue if motorists would adhere to the legal following distance, because there would always be a gap in which to merge.  This is one reason why I think following distance should be enforced much more often.

[plain]

This is I-64 EB as it merges onto I-95 SB.
https://goo.gl/maps/N3rEK9P2usM2

The problem here is not only the onramp's left lane merge itself, but the right onramp lane becomes an exit only lane for the upcoming Boulevard exit, which is less than a 1/2 mile away. This creates a serious weaving issue for both interstates and a major chokepoint for I-64 EB traffic during both rush hours.

[johndoe]

Brandon,  I'm referring to either the mainline lane vs the ramp lane yielding , not the two ramp lanes.

Drivers shouldn't assume they have the right of way at any merge.  If one is explicitly told to yield, then, yes, one should yield.  But in a merge where neither side has a yield sign, drivers should have enough sense to act courteously and cautiously when they see that two vehicles are trying to occupy the same space ahead (slow down a little if necessary, but don't stop if you can avoid it, and don't play chicken).  At any rate, this would rarely be an issue if motorists would adhere to the legal following distance, because there would always be a gap in which to merge.  This is one reason why I think following distance should be enforced much more often.

Perhaps this isn't correct legally, but it seems to me anyone in a lane that's ending IS being told to yield.  It's their responsibility to find a gap.  Sure it's nice to let someone in, but someone in the mainline lane better not be panic-braking.  IMO the inside merge leads to more of these "chicken" moments since drivers on both sides aren't clear on priority.  (Like walking toward someone and you both go the same way...both accelerate, both decelerate, etc)   I suppose one problem with a legal following distance is that it would hurt capacity.

I'm also curious how these merges are modeled in HCM or, more importantly, microsimulation.  Does anyone know if this can be done, or how accurate it is?

[Ned Weasel]

I suppose one problem with a legal following distance is that it would hurt capacity.

I'm pretty sure braking a lot to avoid hitting the car ahead when one is following too closely, hurts capacity.  I'm pretty sure drivers not leaving a big enough gap to be overtaken, hurts capacity.  I'm pretty sure rear-end collisions resulting from following too closely, hurt capacity.

[J N Winkler]

I'm also curious how these merges are modeled in HCM or, more importantly, microsimulation.  Does anyone know if this can be done, or how accurate it is?

It can be done:  something something Markov chains.  Leclerq has apparently been working on capacity at merges at Georgia Tech for the last five years or so, and this 2011 paper is useful partly for the bibliography and partly for the account of how the Newell-Daganzo model works:

http://trafficlab.ce.gatech.edu/sites/default/files/files/documents/leclercq_isttt.pdf

However, I'm not sure work has been done that addresses the capacity and safety profiles of various merge striping patterns.  I suspect safety is partly LOS-dependent.

[kphoger]

it seems to me anyone in a lane that's ending IS being told to yield.  It's their responsibility to find a gap.  Sure it's nice to let someone in, but someone in the mainline lane better not be panic-braking. 

It's not always clear at a merge which lane is ending and which lane is continuing.  Sometimes the design is more of a "form one lane" type thing.  Whenever I encounter situations like this, I assume traffic should simply take turns, and I accelerate or decelerate appropriately.

In the Richmond example posted by plain above (https://goo.gl/maps/N3rEK9P2usM2), it's not obvious–or even reasonable to say–whether I-64's lane or I-95's lane is ending.

[johndoe]

I suppose one problem with a legal following distance is that it would hurt capacity.

I'm pretty sure braking a lot to avoid hitting the car ahead when one is following too closely, hurts capacity.  I'm pretty sure drivers not leaving a big enough gap to be overtaken, hurts capacity.  I'm pretty sure rear-end collisions resulting from following too closely, hurt capacity.

Like the braking that would be required to maintain some arbitrary distance?  People can judge for themselves an appropriate gap and densities will adjust accordingly around merge areas.

Thanks JN.

Kphoger: to me this lack of clarity is exactly why I cannot imagine a scenario where the inside merge is a good option.  I can't imagine how a parallel acceleration lane followed by a taper wouldn't  always be more clear.

[Ned Weasel]

I suppose one problem with a legal following distance is that it would hurt capacity.

I'm pretty sure braking a lot to avoid hitting the car ahead when one is following too closely, hurts capacity.  I'm pretty sure drivers not leaving a big enough gap to be overtaken, hurts capacity.  I'm pretty sure rear-end collisions resulting from following too closely, hurt capacity.

Like the braking that would be required to maintain some arbitrary distance?

No, because people have to slow down more suddenly when they're following too closely.  When maintaining the correct following distance, one can slow down gradually to maintain the gap when the car ahead slows down gradually.  When following too closely, the car doing the following has to slow down more quickly than the car ahead.  This is why, if you follow someone too closely, it increases your chances of being rear-ended.  You can probably brake fast enough to avoid hitting the car in front of you, but when you have to brake very suddenly just to avoid hitting the car in front of you, because you were following too closely, and someone else is following you too closely, guess what's likely to happen?

People can judge for themselves an appropriate gap and densities will adjust accordingly around merge areas.

Except they too often don't.  Too often, people just follow about as closely as they can without hitting the car ahead of them.  Two seconds is a good rule of thumb (for passenger cars on dry pavement), but how often do you really see people leave that much space?  People too often don't even leave one second of following distance.  At any rate, people are supposed to leave enough of a gap to be overtaken, but how often do you really see them make an earnest effort to do that?  If one is just leaving enough space to avoid hitting the car ahead, that's not enough of a gap to be overtaken.  And if a driver isn't leaving enough of a gap to be overtaken, it hurts merge situations, because it reduces the availability of gaps in which people can safely merge.

[johndoe]

Strident,

I don't really want to hijack this thread (especially since inside merges are such an interesting feature), so here are my last thoughts on this: 

You're right that some people accelerate and decelerate too much to make me comfortable.  I'm more of a bigger-headway guy myself.  But I don't see how changing (or cracking down on?) the law would improve situations.  For one, I'm not sure how headways could be enforced by police.  And even if they could, you mentioned how a constant value doesn't exist.  (Is it raining? Are my brakes poor?  Am I in the middle of nowhere?  Is there a ramp ahead with 10 cars merging?) Plus, I'm pretty sure we don't need to give police any more reason to pull over drivers in peak periods 

VISSIM, the microsimulation program used by many DOT, uses a car-following model called Wiedemann 99.  The default value for headway (CC1) is 0.9 seconds.  Here's a snippet from the WSDOT VISSIM protocol (http://www.wsdot.wa.gov/NR/rdonlyres/378BEAC9-FE26-4EDA-AA1F-B3A55F9C532F/0/VissimProtocol.pdf)
A CC1 of 2 seconds isn't even on the chart because it's so unrealistic, but I think you can "guess what's likely to happen?" to flow rates (and upstream congestion).

Another cool site related to that car-following model is http://w99demo.com/  You can see the action of all the vehicles (accel, follow, decel) and play with differing parameters.

[kphoger]

For one, I'm not sure how headways could be enforced by police. 

I was under the impression that the highway patrol cars in Germany were equipped with cameras and the technology needed to measure following distance, and that they are therefore able to issue valid citations for too-close following distance.

[Ned Weasel]

Strident,

I don't really want to hijack this thread (especially since inside merges are such an interesting feature), so here are my last thoughts on this: 

You're right that some people accelerate and decelerate too much to make me comfortable.  I'm more of a bigger-headway guy myself.  But I don't see how changing (or cracking down on?) the law would improve situations.  For one, I'm not sure how headways could be enforced by police.  And even if they could, you mentioned how a constant value doesn't exist.  (Is it raining? Are my brakes poor?  Am I in the middle of nowhere?  Is there a ramp ahead with 10 cars merging?) Plus, I'm pretty sure we don't need to give police any more reason to pull over drivers in peak periods 

VISSIM, the microsimulation program used by many DOT, uses a car-following model called Wiedemann 99.  The default value for headway (CC1) is 0.9 seconds.  Here's a snippet from the WSDOT VISSIM protocol (http://www.wsdot.wa.gov/NR/rdonlyres/378BEAC9-FE26-4EDA-AA1F-B3A55F9C532F/0/VissimProtocol.pdf)

(Image snipped to save space.)

A CC1 of 2 seconds isn't even on the chart because it's so unrealistic, but I think you can "guess what's likely to happen?" to flow rates (and upstream congestion).

Another cool site related to that car-following model is http://w99demo.com/  You can see the action of all the vehicles (accel, follow, decel) and play with differing parameters.

This article, posted in an earlier thread, is rather informative and relevant: http://www.wsj.com/articles/one-driver-can-prevent-a-traffic-jam-1476204858

[UCFKnights]

For one, I'm not sure how headways could be enforced by police.  And even if they could, you mentioned how a constant value doesn't exist.  (Is it raining? Are my brakes poor?  Am I in the middle of nowhere?  Is there a ramp ahead with 10 cars merging?) Plus, I'm pretty sure we don't need to give police any more reason to pull over drivers in peak periods 

Around here they do occasionally pull people over for their following distance and being obnoxious letting people in. One of my buddys actually got a ticket for it. Its done totally by eye, arbitrarily. Personally, I'd rather the cops pull people over during peak periods when a couple of assholes are ruining it for the rest of us vs non-peak periods where someone being an asshole is easily avoided by giving them their space.

[Revive 755]

The vehicles merging onto the freeway from the lane that merges, should either yield or speed up while finding their gap in traffic.  Those in the lane that gets its own new lane, can do as they please.

Sounds good in theory, does not work well in practice with either large trucks or those drivers who refuse to accelerate until already completely merged onto the freeway.  The latter case usually gets exacerbated by other drivers cutting across the painted gore and instead of waiting for the merge.

The ramp from NB I-355 to EB I-88 is a nice example:  a nice, long approach, leading up to a nice long inside lane merge as you move forward, allowing plenty of time for drivers to assess and react.

https://www.google.com/maps/@41.809586,-88.047163,3a,75y,87.43h,91.13t/data=!3m7!1e1!3m5!1sznjCGERmYU7nUiOlorh_0A!2e0!6s%2F%2Fgeo2.ggpht.com%2Fcbk%3Fpanoid%3DznjCGERmYU7nUiOlorh_0A%26output%3Dthumbnail%26cb_client%3Dmaps_sv.tactile.gps%26thumb%3D2%26w%3D203%26h%3D100%26yaw%3D328.6588%26pitch%3D0%26thumbfov%3D100!7i13312!8i6656

That one does not work very well IMHO; there's usually too much traffic on EB I-88 trying to position for the ramp to NB I-355.  Additionally, given the length of the painted merge gore, there is only about ~250 feet (around 2.5 seconds at what is supposed to be the speed limit on the tollway) for the inside ramp lane and outside lane of I-88 for two vehicles to decide who will actually give way.

[paulthemapguy]

The ramp from NB I-355 to EB I-88 is a nice example:  a nice, long approach, leading up to a nice long inside lane merge as you move forward, allowing plenty of time for drivers to assess and react.

https://www.google.com/maps/@41.809586,-88.047163,3a,75y,87.43h,91.13t/data=!3m7!1e1!3m5!1sznjCGERmYU7nUiOlorh_0A!2e0!6s%2F%2Fgeo2.ggpht.com%2Fcbk%3Fpanoid%3DznjCGERmYU7nUiOlorh_0A%26output%3Dthumbnail%26cb_client%3Dmaps_sv.tactile.gps%26thumb%3D2%26w%3D203%26h%3D100%26yaw%3D328.6588%26pitch%3D0%26thumbfov%3D100!7i13312!8i6656

That one does not work very well IMHO; there's usually too much traffic on EB I-88 trying to position for the ramp to NB I-355.  Additionally, given the length of the painted merge gore, there is only about ~250 feet (around 2.5 seconds at what is supposed to be the speed limit on the tollway) for the inside ramp lane and outside lane of I-88 for two vehicles to decide who will actually give way.

Well, it's better than some of the single-lane entrance ramp convergences that are out there.  I like that there's a considerable amount of side-by-side travel before the point of convergence.  That is more important for anticipating traffic than the gore zone itself. (imo of course)

https://goo.gl/maps/UAE6NU8F2Vv

...But I find it utterly unacceptable that this sign doesn't have a yellow "EXIT V ONLY" arrow

https://goo.gl/maps/xAsVwWXU3z32

[frankenroad]

Ohio loves these, especially in the Cleveland area. There's one at either end of the I-271/I-480 concurrency, 3 at the I-90/I-271 interchange and many more in that area. In Columbus, there's one at I-270/I-670.

Also in Cincinnati where 74 merges into 75 South.