Speed limit reduction towards cities encourage traffic jams?

[colinstu]

I find it interesting that entering milwaukee from I-94 east-bound goes from 65mph to 55mph (around 10-15mi away from downtown), to 50mph (5-10mi away). As traffic gets closer to the city, it's suggested that traffic slows down. This means that people are basically being rushed towards a possible jam. Wouldn't it be better to have all the speeds the same (all 55 for example)? Or speed up the closer one is to the city / traffic area?

The same can be said for I-43 south-bound towards downtown. Same story with 65->55->50->45 for a curve->50. I find myself cruising the fast speed limit and then ending up at a jam where everyone is required to slow down / is already slow due to not enough capacity. If me and other drivers were to instead drive 5-10mph slower, maybe the jam would have a chance to mitigate before we arrived and stomped on our breaks / slowed down for low speed limits.

This is basically a large-scale version of coasting behind cars during a jam instead of parking right on someone's tail and doing the stop&go BS.

[NE2]

No. If I'm not mistaken, slower speeds mean more throughput because of less following distance.

[Alps]

Traffic engineer here. The slowest speed limit has the lowest throughput. Smaller following distance is offset by the longer amount of time it takes each vehicle to pass a given point. The real issue is that people will go whatever speed is comfortable. The 45 mph curve will cause a jam no matter what the speed limits are leading up to it, because people will go 60-65 until they see the curve and then HOLYSHIT45NOW.

[NE2]

http://www.wsdot.wa.gov/smarterhighways/vsl.htm
"Ideally, approaching traffic will slow down and pass through the problem area at a slower but more consistent speed reducing stop and go traffic."

I suppose it makes the average speed higher by reducing the amount of wicked slow traffic?

[Alps]

http://www.wsdot.wa.gov/smarterhighways/vsl.htm
"Ideally, approaching traffic will slow down and pass through the problem area at a slower but more consistent speed reducing stop and go traffic."

I suppose it makes the average speed higher by reducing the amount of wicked slow traffic?

Big difference: When you have variable limits and they suddenly show lower, people are likely to obey them because they recognize there must be an unusual condition. When you have static limits, people just go the speed they know once they've familiarized themselves with the road. VSL are absolutely effective.

[NE2]

Except here you're talking about actual speed being lower:

Traffic engineer here. The slowest speed limit has the lowest throughput. Smaller following distance is offset by the longer amount of time it takes each vehicle to pass a given point.

[Alps]

Except here you're talking about actual speed being lower:

Traffic engineer here. The slowest speed limit has the lowest throughput. Smaller following distance is offset by the longer amount of time it takes each vehicle to pass a given point.

No. If I'm not mistaken, slower speeds mean more throughput because of less following distance.

You are wrong.

[NE2]

I accept that. But now I'm really confused about what I was trying to say.

[kphoger]

Let's all have a nice scream.

[1995hoo]

http://www.wsdot.wa.gov/smarterhighways/vsl.htm
"Ideally, approaching traffic will slow down and pass through the problem area at a slower but more consistent speed reducing stop and go traffic."

I suppose it makes the average speed higher by reducing the amount of wicked slow traffic?

Big difference: When you have variable limits and they suddenly show lower, people are likely to obey them because they recognize there must be an unusual condition. When you have static limits, people just go the speed they know once they've familiarized themselves with the road. VSL are absolutely effective.

That's the theory, at least, but it didn't work when it was tried here in Northern Virginia through the Wilson Bridge—related work zone on the Beltway. People already ignored the regular 55-mph speed limit and they also ignored the variable speed limit that was in place for two or three years. I've long wondered if it might be different if the "normal" speed limit were set at a number most drivers consider reasonable; I've also long wondered if part of the problem was that the variable limits began well to the west of the area most people viewed as the work zone. (No doubt the intent was to try to get people to slow down in advance of the work zone, but I strongly suspect most drivers viewed it the same way they view it when highway departments post lengthy "work zones" in which no work is apparent to any observer.)

[cpzilliacus]

Traffic engineer here. The slowest speed limit has the lowest throughput. Smaller following distance is offset by the longer amount of time it takes each vehicle to pass a given point. The real issue is that people will go whatever speed is comfortable. The 45 mph curve will cause a jam no matter what the speed limits are leading up to it, because people will go 60-65 until they see the curve and then HOLYSHIT45NOW.

Related to this, the District of Columbia has automated speed enforcement devices along its (few) miles of limited-access highways (most of them look like recycled traffic signal control cabinets, because that's what they are). 

D.C. posts speed limits that are way below the 85th percentile (currently the highest is 50 MPH), and locals invariably apply the brakes as they approach one of those devices.

Even worse if a speeder triggers one of them at night, since that results in an extremely bright strobe light going off, which causes anyone who sees it to hit the brake pedal.

[agentsteel53]

should we be re-evaluating how brake lights work?

it is a challenge to discern the difference between a car in front of me "adjusting velocity down 3mph by hitting the brakes" vs "slamming on the brakes because the roadway is missing ahead".

[1995hoo]

....

Related to this, the District of Columbia has automated speed enforcement devices along its (few) miles of limited-access highways (most of them look like recycled traffic signal control cabinets, because that's what they are). 

D.C. posts speed limits that are way below the 85th percentile (currently the highest is 50 MPH), and locals invariably apply the brakes as they approach one of those devices.

....

Indeed, and it always amuses me when you get someone from out of town on the road. All of us will be bombing along at 65 to 70 and then we abruptly slow to between 45 and 50. The non-local drivers usually get pissed, start trying to change lanes, weaving back and forth to show displeasure, etc......if you're not a local driver, and you see that all the local drivers abruptly reduced their speed, wouldn't you think there's probably a good reason why they're doing it?

should we be re-evaluating how brake lights work?

it is a challenge to discern the difference between a car in front of me "adjusting velocity down 3mph by hitting the brakes" vs "slamming on the brakes because the roadway is missing ahead".

I seem to recall reading that Mercedes was given a special permit to test a brake light system in which slamming on the brakes causes the brake lights to flash rapidly. I haven't heard anything about the results. I know when I'm at the tail end of a traffic jam on a high-speed road I'll usually turn on my hazard flashers as a way of telling approaching traffic that I'm stopped, not just slowing down.

[colinstu]

I've thought about it in the past, variable intensity (or number of) brake lights depending on the pressure on the pedal.

Regular pressure turning on all three brake lights, more pressure turning on 2 lights, etc. Like new Ford Mustangs that have the sequential turn signals (each side is broken into 3 chunks. only 1 on each side would be lit with light braking, with all of them on during extreme braking).

I've also noticed some ambulances that have strobing brake lights. When pressed, they light up in varying intensities ("on" and "holy shit strobing on"). Maybe requiring this whenever a car's ABS system is engaged.

3...2...1... until this thread turns into computer-controlled braking systems that sense speed and distance of cars in front of them.

Not where I wanted this to all go. I'm still not convinced that having fast speed limits towards a slow area helps anything (even if following distance changes. People in Wisconsin follow 1 second behind if you're doing 70mph or 40mph), and that "no one will obey variable speed limits" is a valid excuse for not using such a system.

[NE2]

I think the idea is that the speed limit reduction is intended to reduce traffic jams that exist due to other factors. Everyone's going to be going slowly when they reach the jam, but if you can spread out the slow traffic you may be able to increase the average speed or eliminate the stop-and-go section entirely. It seems clear that, even from the standpoint of the individual driver, steady flow at 45 is preferable to steady at 65 for half the distance and then stop-and-go for the other half, but even a strapping young genius like me got confused thinking about it. And don't forget about the eejits who think the gummint is just trying to get them to slow down for no reason.

[agentsteel53]

I think the idea is that the speed limit reduction is intended to reduce traffic jams that exist due to other factors. Everyone's going to be going slowly when they reach the jam, but if you can spread out the slow traffic you may be able to increase the average speed or eliminate the stop-and-go section entirely. It seems clear that, even from the standpoint of the individual driver, steady flow at 45 is preferable to steady at 65 for half the distance and then stop-and-go for the other half, but even a strapping young genius like me got confused thinking about it. And don't forget about the eejits who think the gummint is just trying to get them to slow down for no reason.

if we have proven that the entire segment can handle 45 steady flow, and the first half of the segment can handle 65 steady flow - why have the stop/go section at all?  how can we get people to slow down from 65 to 45, as opposed to a dead stop?

[NE2]

if we have proven that the entire segment can handle 45 steady flow, and the first half of the segment can handle 65 steady flow - why have the stop/go section at all?  how can we get people to slow down from 65 to 45, as opposed to a dead stop?

By not having traffic stopped in front of those people

Traffic jams are complicated things. You can't just clear them up by saying "OK everybody, go 45 now".

[agentsteel53]

By not having traffic stopped in front of those people

Traffic jams are complicated things. You can't just clear them up by saying "OK everybody, go 45 now".

the question is, what gets them stopped?  "person N is stopped because person N-1 is stopped" is great inductive reasoning, until you reach the base case - there is, by definition, someone who is at the front who is stopped, even though the person in front of him did not stop.

this person must be found, and shot.

[Crazy Volvo Guy]

Traffic engineer here. The slowest speed limit has the lowest throughput. Smaller following distance is offset by the longer amount of time it takes each vehicle to pass a given point. The real issue is that people will go whatever speed is comfortable. The 45 mph curve will cause a jam no matter what the speed limits are leading up to it, because people will go 60-65 until they see the curve and then HOLYSHIT45NOW.

The sad part is, most cars can easily exceed those posted curve advisory speeds, but their drivers won't.  Nine times out of ten, those advisory speeds are appropriate for me, and I drive this.

[colinstu]

By not having traffic stopped in front of those people

Traffic jams are complicated things. You can't just clear them up by saying "OK everybody, go 45 now".

the question is, what gets them stopped?  "person N is stopped because person N-1 is stopped" is great inductive reasoning, until you reach the base case - there is, by definition, someone who is at the front who is stopped, even though the person in front of him did not stop.

this person must be found, and shot.

It can be caused by someone over reacting and braking too hard... to genuine poor design of roads (too many people being forced in too little lanes / too quickly).

I think that's a big part of the cause that develops on I-43 southbound toward the Marquette interchange. There are four problem areas I've found.

First there is a tight curve with signs that flash demanding people to slow down. Once people have slowed down a little, they are now facing the setting sun. Even if your visor is down and windshield clear, visibility is very poor. Not expecting the sun there? Drivers will be blinded. Thirdly there is traffic merging from Fond Du Lac Ave weaving with traffic wanting to exit toward I-94 West and I-794. And Fourthly (and the biggest cause of slowdowns on the 1-3 left lanes) is all the CRAP that has to merge together in I-94/43 southbound. 8 lanes of traffic merging down into 3 (not including 1 lane that is an exit-only). I43 goes from 3->2 lanes, Wisconsin Ave then comes in and has to merge too into those 2 lanes, I-94 East comes in as 2->1 and adds a lane to the I43/Wisconsin traffic, and then I-794 comes in as 2->1 lane and ends up in the exit-only lane. It's this monster that causes the most slow downs.

Ugh.

[Crazy Volvo Guy]

And don't forget about the eejits who think the gummint is just trying to get them to slow down for no reason.

I wouldn't say they're all idiots, they're just jaded.  For all too many years (to this day in some places) speed limits were set low and enforced heavily for no reason other than to increase revenue for the city/county/state and profit for the insurance companies (as they were and generally still are in bed with each other.)  The result today is widespread disrespect toward and ignorance of low speed limits.

[NE2]

I did a bit more thinking, and came up with some conclusions:

1) Assuming 2 second following distance, speed does not affect throughput by much, as one vehicle per lane passes a point every two seconds no matter what the speed. (Faster speeds mean that the time between front and rear bumper decreases slightly, but that probably doesn't matter much at normal free-flow speeds.)
1a) If people don't follow safely, that's almost certainly the reason for many traffic jams - the person that Mr. Steel wants to shoot was following someone too closely who slowed a bit to let someone in, or because a hill is beginning, and had this person not been following too closely he could have merely eased off on the gas without hitting the brakes.
2) Throughput and capacity are two different things. A parking lot has a lot of capacity but almost zero throughput. When dealing with commuting patterns, capacity may be more important, since x people need to be somewhere by 9, as opposed to x people needing to be through the area between 8 and 9.
2a) Slower speeds increase capacity, since following distance decreases.
3) Given a certain amount of traffic on the road, slowing it down will increase following distance, reducing the chance of new jams.
3a) Therefore, slowing traffic upstream from a congestion point will decrease throughput into the point, resulting in better flow and perhaps even clearing up existing problems.
I'm probably missing something here.

[kphoger]

1a) If people don't follow safely, that's almost certainly the reason for many traffic jams - the person that Mr. Steel wants to shoot was following someone too closely who slowed a bit to let someone in, or because a hill is beginning, and had this person not been following too closely he could have merely eased off on the gas without hitting the brakes.

This was my immediate reaction to agentsteel53's "base case" hypothesis.  It doesn't necessarily follow that someone stopped with no one in front stopping.  All it takes is someone to slow down to, say, 20 mph–which may well be legitimate for any number of reasons.  The person following sees the brake lights and slows to 18 mph.  The person following him plans to slow only to 22, realizes his error, and steps hard on the brakes down to 12 mph, then accelerates slightly to match pace.  The person following him sees him slam on the brakes and slows to 10 mph.  And so on.  Meanwhile, people far in back of them and those in adjacent lanes see a sea of brake lights and sudden movements, and slow down, perhaps more than necessary, as a matter of defensive driving.  And the process multiplies.  Much of this, but probably not all, could be corrected by giving proper following distance.

It should be noted that sudden lane changes during heavy traffic reduce throughput–likely due to differences in speed encountered in each lane, i.e. the one that was flowing slightly better until Mister Hot Dog entered it from a dead stop.  California found this out while studying the effect of adding HOV lanes:  even though the HOV lane had far fewer cars in it than the other lanes (thereby increasing the number of cars in each of the other lanes), throughput still increased–which fact was determined/theorized to be because of fewer lane changes due to fewer regular lanes.

[Crazy Volvo Guy]

1a) If people don't follow safely, that's almost certainly the reason for many traffic jams - the person that Mr. Steel wants to shoot was following someone too closely who slowed a bit to let someone in, or because a hill is beginning, and had this person not been following too closely he could have merely eased off on the gas without hitting the brakes.

This is probably the most common cause, all things considered.  Nobody understands/cares what a proper following distance is, anymore.  (Did they ever?)  Seeing how all I do is drive the highways anymore, I see this phenomenon first hand, every day of my life.   People just bunch up when  the flow slows down, barely leaving a car-length between each other; whether it's behind someone going slow in the left lane, or when all the lanes slow down.

[Kacie Jane]

1a) If people don't follow safely, that's almost certainly the reason for many traffic jams - the person that Mr. Steel wants to shoot was following someone too closely who slowed a bit to let someone in, or because a hill is beginning, and had this person not been following too closely he could have merely eased off on the gas without hitting the brakes.

This is probably the most common cause, all things considered.  Nobody understands/cares what a proper following distance is, anymore.  (Did they ever?)  Seeing how all I do is drive the highways anymore, I see this phenomenon first hand, every day of my life.   People just bunch up when  the flow slows down, barely leaving a car-length between each other; whether it's behind someone going slow in the left lane, or when all the lanes slow down.

Or if you do leave a car length or two, someone takes it as an opportunity to change lanes into that spot for no discernible reason, thus negating your efforts.