Slashdot Mirror
Chaos and Your Everyday Traffic Jam
An anonymous reader writes "What causes these mysterious traffic jams that continually appear throughout the day for no reason whatsoever? Is it simply the fact that most people just don't have a clue how to drive? That's very possible, and in reality there are so many variables involved in something like a traffic jam. But is it possible that the entire traffic jam could be both the continuing and end result of a chain reaction set in motion by a single driver who was in too much of a hurry?"
I fly light planes. Major roads, when VFR, are very good landmarks.
... then the drivers behind (probably following far too close) brake a bit harder, and the drivers behind them brake a bit harder still. The adjacent lanes, in seeing one lane suddenly slow go 'whoa', and someone also brakes in that lane. Pretty soon, just from one person braking a little bit - the braking has propagated down the road with greater and greater severity until one of two things happens: usually, the traffic comes to a standstill, and you get a self-sustaining standing wave of stopped traffic until the amount of traffic on the road reduces to the extent there are fewer cars joining the wave than are leaving. This can take HOURS, especially on the M6 in England. The second thing that may happen in this cascading braking severity is that someone runs into the back of the other. Then chaos ensues for most of the day.
Quite often when it is very busy, you can see a standing wave in the traffic - there's an area where all the cars are stopped - but there is NO obstruction at all. The cars are filling the 'standing wave' from the back as quickly as cars at the front are leaving it - so it becomes self-sustaining.
When the road is full to capacity, moving at 70 mph, all it takes is one person to jab their brakes
The other problem is lorries (large trucks) overtaking lorries with a speed differential of 0.5 mph. It takes them several minutes to get past because they are both speed limited within 0.5 mph of each other, meaning the inside two lanes are 56mph, and the outside lane is 70mph+. When a frustrated driver pulls out into the outside lane after being stuck behind a lorry for "too long", they cause one of the outside lane drivers to brake down to 56 mph quite suddenly. This can easily get the 'braking cascade' started, and before you know it - you have a standing wave traffic jam with no actual obstruction (other than the standing wave itself).
Usually then what happens, is the opposite direction traffic, seeing the stoppage rubber necks for the possible accident. An inattentive driver looking at the other side of the road finally looks back in front and realises he's about to ram a truck in the rear and slams on the brakes. The driver behind him following far to closely has to brake even harder - and there's either a shunt or if they are lucky, ANOTHER standing wave traffic jam starts on this side of the road too.
It's fascinating to watch from the air. Frustrating to be in when driving.
Oolite: Elite-like game. For Mac, Linux and Windows
See this Science Hobbyist article from January 1998. It's long and detailed, and suggests practical steps individual drivers can take for breaking up (or causing!) traffic jams. Yes, dear readers, this is a nine-year-old dupe.
This post expresses my opinion, not that of my employer. And yes, IAAL.
rar42 (626382) sez:
> I'm inclined to compare roads to shared medium Ethernet. As the
> traffic builds up you get more 'collisions' and both systems
> have collision detection built-in. With Ethernet, as the 'traffic'
> builds to about 40% of the theoretical capacity, collisions
> become the norm
You're pretty much completely wrong, and the last quoted line sums up why.
Collisions are not the norm in traffic jams.
Traffic jams happen due to the ripple effect from cumulative reaction time
delays in response to changes in traffic. The effect accumulates until there
is so much loss of speed that people drive closer together. Then when they
have to react, they react more abruptly, and that causes yet a stronger ripple
effect.
Packets collide, cars don't. Cars change speed, packets don't.
Well, OK, sometimes cars do collide. But it's not the collision itself that
causes the traffic jam, it's the bottleneck in the right of way and/or the
rubberneckers.
If people could and would simply maintain the 2 second following distance
no matter what speed, when the fewer traffic jams did occur, they would resolve
themselves much more quickly. But just try telling the person 500 cars back to
just sit still for 10 minutes. They'd probably want to punch you, and they'll
still insist on driving stopgostopgostogostopgo despite the fact that doing so
means they'll be doing it for several times longer than just waiting.
90% of drivers think they're better than average.
90% of drivers are below average drivers.
So I give free driving lessons.
Like braking suddenly for tailgaters.
"I may be synthetic, but I'm not stupid." -- Bishop 341-B
The two-second gap is typically recommended for 70 MPH speeds. It is dramatically smaller for lower speeds. It is also completely unrealistic at any speed. A two-second gap at 70 MPH is about 315 feet, or about 24 car lengths. At 70 MPH the assumption is that the average person requires about 2/5th of a second to react, which equates to just under 70 feet. That leaves 247 feet for braking. The NHSTA pegs the average 70 MPH braking distance for modern cars around 170 feet. I personally suspect it's much lower these days, and with panic braking plus ABS it would be far, far lower. But give them the benefit of the doubt (though god only knows why) and call it 160 feet. That leaves you 87 feet, so the two-second gap is overly conservative by a factor of about 25%.
... 16 cars per mile and the highway would look positively empty.
BUT -- and here is where it gets stupid for real-world conditions -- that braking scenario assumes that you must stop within that two-second gap. Think about this: the only way that would matter is if there is an immobile object two seconds ahead of you. You're driving along, then mysteriously, 315 feet in front of you, something is stopped dead. What are the actual chances of this happening to any responsible, alert driver doing 70 MPH? Very small. In fact what will happen is that the car ahead begins to slow, and you burn your 0.4 second reaction time (which I also think is unrealistically high), then you begin to slow in concert with the car ahead. It is obviously impossible to derive any specific numbers for the rates at which this happens as they'll be random, but it certainly doesn't equate to a 70MPH-to-zero panic stop in a limited space.
In any case, his figure of 120 cars per minute is probably a lot closer to reality than anything provided for by the 2-second rule, which is a 24-car-length gap -- have you ever seen a busy highway where anyone was maintaining a 24-car-length gap? Would it even be possible to actually estimate and maintain this at 70 MPH? That's about 1/16th of a mile
Slashdot quality declines as the number of hot grits posts decreases. - Provolt's Law, Apr-09-2005
In our area, there is a twice-daily traffic jam that has been understood for years, but fixing the road to take away the problem would be ungodly expensive.
There is, actually, nothing technically wrong with the road. The road in question is I-87 (the Northway), and the pinch point is where it crosses the Mohawk river. The Twin Bridges have a slightly narrower shoulder than the highway leading up to them in either direction, but the shoulder, on both sides of each bridge, is still every bit as wide as any of the three lanes going in either direction.
Compounding the problem is that the bridges are (hope this is the right term) truss bridges. There are two convex bowed beams that go over each side of each bridge, and a construct of triangular trusses between them. These are the reason why a change would be ungodly expensive, because you would have to rebuild the bridges.
Anyway, people come to the bridges and slow down because they perceive that the road has gotten narrower, while failing to perceive that this fact is irrelevant. This slowing down leads to the accordion effect that was described in TFA, where successive cars have to apply more and more braking in order not to hit the car in front of them. By the time you are a mile north of the bridge in the mornings (south in evenings), traffic is basically stopped.
The construct that causes all of this trouble can be seen here (along with some Google wierdness in the construction of the image).
www.wavefront-av.com