warren@ihnss.UUCP (06/13/83)
Having spent another wasted hour on the "Tri-state crawlway" over the weekend, I got to wondering why traffic speeds on overcrowded freeways tends to oscilate, stopping and accelerating, rather than maintaining a constant though slower speed. Such oscillations occur even when there is no entering or exitting traffic, as is true of long sections of this particular road. Given the annoyance of stop and go driving in a standard shift car, I would presume that at least some of the people on the road would be trying to damp out the speed variations. Is there some physical law that dictates this behavior, or is it a result of poor driving practices. Given an appropriate set of mathematical formulas describing the behavior of freeway traffic, I wonder if it would be possible to improve the flow of traffic through bottlenecks by minor modifications of the roadway. I have seen a lot of this sort of analysis on traffic lights, but none on clogged freeways. Always looking for elegant solutions to real-world problems ... -- Warren Montgomery ihnss!warren IH x2494
dje@5941ux.UUCP (06/13/83)
Having to commute up and down NJ Route 287 every day during rush hour, I too
wonder about the reasons for stop-and-go traffic. One possible reason is that
a car will not always match the speed of the one in front of it. If one
vehicle is traveling at 40 mph and the one behind it slows to 30 mph, then all
cars following in that lane have to slow to 30 mph (or else pack that much
tighter, which also slows things down). If traffic is sufficiently heavy, this
kind of chain reaction may continue until one driver approaching the car in
front has to stop before hitting it. It is similar with cars changing lanes:
when a car enters a lane, the car behind it has to crowd in closer or slow down
to make room.
If mathematical solutions are hard to come by, a simulation model might also
be helpful in providing insight into the process.
Dave Ellis / Bell Labs, Piscataway NJ
...!{ariel,lime}!houti!hogpc!houxm!5941ux!dje
...!harpo!npoiv!npois!houxm!5941ux!djemat@hou5e.UUCP (06/14/83)
Regarding the question of waves through traffic, and driving habits: This effect is documented, and is, I believe, well understood as a topic in civil engineering. Regarding especially driving habits -- we are poorly trained when taught to drive. As an example, consider two situations. One is the typical heavy traffic situation. We tend to gun the gas to accellerate quickly, and end up braking when the driver ahead comes to his peak speed. The person behind is worse off than we are. The problem is often compounded by automatic transmissions that upshift with the engine just above idle and suddenly lurch the car forward faster. In really heavy traffic (say 5-20 mph, or even up to 35 mph) I pull my slushpot down to 2 or 1, keeping the car moving faster if need be with a little gas. As a side benefit, the chargine system in my car, which is none to zoftig, is allowed to turn at a speed where it is almost fullt effective. Also, if the driver ahead of me is tailgating, I keep about 2-1/2 seconds back so that I won't end up magnifying his quick and sudden but small speed corrections as I drive. If he has plent of room, I will move up to about 1-1/2 seconds behind because he will hot need to stop or slow down suddenly. I also watch two or three cars ahead (at least). The other situation is Red Bank. For those who are not familiar, Front Street runs east-west through RB, crossing first the southbound lanes of Route 35 (a Major Artery). About two blocks later it crosses the northbound lanes of 35, and after another block or so, it intersects Broad street. All of these intersections are controlled by traffic lights. These lights do not appear to be synchronized, so that one usually ends up waiting for several cycles at all three lights. It is important here, as in any city driving, to quickly (but safely) pull out when the light turns green, and pull with a little haste right up to the car in front of you. Most drivers take their sweet time, and as a result, traffic through RB remains backed up for a half-mile or more in each direction on Front Street. Note that on the crowded highway, where patience and reserve would speed traffic, most people drive aggressively, driving the stakes into the heart of their timetables. In Red Bank, or other city driving, where a certain amount of informed assertiveness would speed traffic vastly, most people figure that there is nothing they can do, and as a result screw up the folk behind. Mark Terribile Duke of DeNet
/a3/c5215a/weheh/user@mhuxa.UUCP (06/14/83)
A physical phenomenon that I have witnessed commuting to and from
work is what I call the "Venturi effect" (after the physical phenomenon
of fluids coursing through a pipe): if a car is stalled on a two lane
highway, the fastest lane will usually be the one blocked by the
stalled car.
The answer to why this happens is that as cars from the blocked lane
merge into the unblocked lane, the remaining cars in the blocked lane
speed up to fill the vacuum. Unless the drivers in the unblocked lane
are especially aggressive and unsympathetic, the cars in the blocked
lane will be allowed to merge into the unblocked lane and drive
around the stalled car.
If a car gets stuck in the Holland tunnel, transit police will allow
cars in the blocked lane to pass the stalled car until the tow truck
arrives at the scene. After the tow truck has moved into place
however, the unblocked lane will be given preference.
R. Gordon
mhuxa!wehehkevenb@tekid.UUCP (06/14/83)
A few months ago, SCIENTIFIC AMERICAN ran an article about traffic patterns in city driving. I only glanced at the article, but it looked like a detailed study. Check your library. Keven Boyett
paulsc@tekecs.UUCP (06/14/83)
You might be interested in reading: Herman, Robert, and Keith Gardels. "Vehicular Traffic Flow." Scientific American, pp. 35-43, December 1963. or Wetherell, Charles. "Etudes for Programmers." Prentice-Hall Inc. 1978, pp. 64-66. Actually the Scientific American article is the reference in the Etudes for Programmers chapter. The idea that "if everybody drives more slowly, everybody gets home sooner" [Wetherell78] is an interesting concept. Of course everybody doesn't refer to each individual person, but to everyone put together (i.e. the average). The idea is that traffic flow has some of the same properties as (nearly?) incompressible fluid flow. (At least I don't think of my car as being very compressible.) The Scientific American article shows results from experiments performed in New York using real people in real traffic. I thought it was interesting enough that I performed a computer simulation for a project for a simulation course I took. (I got the idea from the Wetherell book.) The shockwaves set up by one slow driver, or one stall propogate and dissipate very slowly on a busy highway. The busier the highway the more slowly things clear up. Paul H. Scherf P. O. Box 1000 Del. Sta. 61-201 Tektronix Engineering Computing Systems Wilsonville, Oregon, USA UUCP: ...!XXX!teklabs!tekecs!paulsc (where XXX is one of: aat cbosg chico decvax harpo ihnss lbl-unix ogcvax pur-ee reed ssc-vax ucbvax zehntel) CSNET: tekecs!paulsc @ tektronix ARPA: tekecs!paulsc.tektronix @ rand-relay
leichter@yale-com.UUCP (06/16/83)
Traffic engineers do a great deal of analysis of "the physics of traffic". I know nothing about the details, but fundamentally they view traffic in terms of generally quite turbulent fluid flow. I think the bunching up you see is often essentially a series of standing waves. (This is obvious in certain cases, e.g. when there is an accident that slows traffic down - change in impedence, if you will. You see everyone slow down, then speed up - and the effect persists long after the original cause is gone. It only breaks up when the traffic flow gets light enough - average inter-car distance is large enough - so that the standing wave can be disipated. (In heavy traffic, when the cars ahead of you brake, you must brake, too; hence those behind you brake, etc.) The breakup of the wave undoubtedly looks like the high-frequency limit for sound transmission in a gas when you set it up mathematically, For a very basic discussion, see The Amateur Scientist column in the March Scientific American. Now, for those who like HARD problems (this is from an old American Mathe- matical Monthly, if I remember right): Cars A and B, at time t=0, are at rest with B just behind A. Starting at t=0, Car A begins moving with a constant acceleration. Car B follows behind A as closely as possible, subject to the constraint that it must be, say, k "car lengths" behind A per mph of its own speed. (I.e. it maintains a separation linearly proportional to its speed. Problem: express Car B's speed (or positio)as a reasonable function of time. -- Jerry decvax!yale-comix!leichter leichter@yale
jlilien@sdcrdcf.UUCP (06/16/83)
Regarding freeways: Here in Los Angeles, the main creator of snarls on freeways are unmetered intersections (such as Santa Monica/San Diego freeways interchange) and reductions in the number of lanes. However, the oscillating effect can be noticed where these do not occur. I suspect that the problem is that when stupid drivers begin going up a hill, they don't understand that they need to give the engine more throttle or the car will slow down. This, plus the frustrated drivers behind them that try to pass them by changing lanes causes the traffic to slow. Going down the hills has the opposite effect. Since the traffic is very thick (four or more lanes each way as far as the eye can see on a clear day), one such foulup as this could affect traffic three or four mile down the road. Another problem on the freeways here is the impatient driver who is constantly changing lanes to go a little faster. The funny thing is that when several of these types change from one lane to another at about the same time, there is suddenly a lot of space made in the old lane so the traffic looks as if it is moving faster. And so they change again. It is my belief that traffic would flow much better in rush our if drivers would endeavor to maintain steady (if slowed) speed and reduce the lane changes to a minimal. And drivers that can't drive comfortably in such traffic and handle it by slowing down to a crawl should choose other roads to take or other times to drive. There presence on the road at this time is the greatest menace of all. Such is a dream world though; it will never happen! joel
jonab@sdcrdcf.UUCP (06/16/83)
Another possible cause of the ocillation effect is when one driver touches his brakes, this causes other drivers to touch theirs even if they are not in the same lane. So, one over-cautious driver can easily start a traffic jam by jamming on his brakes. This causes others to do the same, and soon everyone has to slow down behind them. I have noticed this happening many times.
rcf@qumix.UUCP (06/18/83)
I have observed that there will be certain interchanges that seem to have slow downs near them regardless of the traffic volume. There was one in particular that seemed inexplicable till I noticed a feeling of visual constriction approaching it -- if that caused everyone to slow down slightly could this be enough to start a backup. Is there anything other than anecdotes to support this idea? Robert Field Qume Corp. San Jose
dafa@ihuxs.UUCP (06/19/83)
Apropos the discussion of traffic jams, here is "The Straight Dope" from Cecil Adams in the Chicago Reader: My friends and I were trapped in the middle of the Santa Monica freeway, unable to move in any direction, when the conversation turned to the cause of our condition. "Why," one friend asked, "does traffic come to a stop on a highway that presumably offers nothing to stop it? We should be able to drive across the country and back without stoppping, except for gas." It sounds like a silly question, but what stops the first car in the daily freeway tie-up? - D. S., nearing th Vermont off-ramp, Los Angeles I don't know much about the Santa Monica freeway, D., but I would venture to say that if you are near the Vermont off-ramp, part of your problem is that you are in the wrong damned state. See if you can work on this. In the meantime you may be interested to know that engineers have devoted considerable study to expressway traffic, and they have concluded that there is a compelling psychological principle that causes the cars to stop, namely the fear of flaming death. Here's what happens. In theory, given the old rule about maintaining one car length ahead of you for each ten miles per hour driving speed, the capacity of a single lane of expressway is 40 cars per minute (2,400 per hour) at 60 MPH. In practice, however, drivers instinctively begin to slow down at loads higher than 25 cars per minute (1,500 per hour). At 33 cars per minute (2,000 per hour), average speed drops to 35 MPH. At this critical juncture, drivers are extremely jumpy, and they will slam on the brakes at the slightest provocation - anything from an accident or a stall to a couple of extra cars trying to merge into traffic at an on-ramp. The first guy slows down a little, the second guy slows down a lot, and the third, fourth, or fifth guy may stop altogether, bringing traffic to a halt. That's why you almost never find smoothly flowing expressway traffic at speeds below 35 MPH - it's usually stop-and-go. It also explains why relatively minor increases in traffic volume, such as those caused by mass transit fare increases, frequently result in chaos on the highways. For this reason, if suburbanites had brains instead of mush, they would always support transit subsidies. As usual, however, you can never get people to understand what is good for them.
otto@ihuxi.UUCP (06/19/83)
Regarding the comment of "visual constriction" slowing traffic down, I have also noticed the same thing. On the expressway from Naperville to Chicago there is one place where traffic is almost always slow. Once this section is passed, traffic speeds up. There is rarely any physical reason for the slowdown, such as a stalled car, but there *is* a psychological reason. This particular stretch of roadway is psychologically constricted, in that for about 1/4 mile there are no shoulders on either the left or right sides. The road feels more cramped to drive on than do the other sections, and apparently many people react to this feeling by slowing down. Once the shoulders reappear on both sides everything feels OK again, and traffic speeds up. George Otto Bell Labs, Indian Hill ----------------------
kk9w@pur-ee.UUCP (06/20/83)
If the traffic situation is viewed as a standing wave, then where are the reflected cars? Guess I'll have to be more careful driving to school the next couple of days. Dave
rh@mit-eddi.UUCP (Randy Haskins) (06/26/83)
There is also the 'action at a distance effect' on freeways: If there is an accident in one direction, there is usually one in the other direction, since some of the yo-yo's going the other way turn their heads to see the accident, and the yo-yo's in front of them SLOW DOWN and turn their head to see the accident. And don't forget the 'tunnel macho-driver effect.' I was coming back to MIT from Logan Airport in a cab, and we were in the Callahan tunnel. Then there was a siren. I'm thinking: "Oh &*$#, now what do we do?" Suprisingly, in a rare act of reasonability for Boston drivers, everyone got over into the right lane while the ambulance went by. My cabbie had nerves of steel, so he cut in behind the ambulance when it came by. Well, not right behind. Another cabbie had already done this, and he was within (no exaggeration) 5 feet of the ambulance. But the second guy was following at about 10 feet, and that was plenty of room for my cabbie to cut in. So we rode through the tunnel (in an ever-growing line of cars) behind the ambulance at about 35-40 mph. Fun, huh? -Randy