Thursday, March 19, 2009
First rule of ant traffic: no overtaking
Ever seen an ant traffic jam? Researchers studying ant traffic are beginning to understand why
In exploring their environment, ants create huge trail systems like motorway networks. Many researchers have remarked that we may have much to learn from the way ant traffic flows along these trails which seem to be free of the jams that plague our roads. After all, it's quite possible that evolution has somehow optimised ant traffic.
But most studies so far have concentrated on two way traffic in which head on encounters between ants dominate the dynamics. We looked at one such study of bidirectional ant traffic not so long ago.
But what of one way traffic? Physicists and motorists alike have long puzzled over the tremendous complexity of behaviour that emerges in one way traffic flow. Who hasn't been stuck in a jam with no apparent cause and that suddenly evaporates for no obvious reason?
Physicists have long known that these effects are closely linked to the density of traffic. Below some traffic density threshold, the flow is always smooth; but creep above this limit and all kinds of traffic chaos ensues.
Now Alexander John at the University of Cologne in Germany and few mates have studied the traffic flow along trails made by Leptogenys processionalis, a more or less average species of ant.
And what they found is quite extraordinary: the average speed of the ants remains constant, regardless of the density of the traffic. There is no transition to a nonlinear flow, at least not in the conditions that this group studied.
Let's put that in perspective. Ant traffic flow is like rush hour traffic on the New Jersey Turnpike travelling bumper-to bumper at the 55 mph.
So what's the secret? John and his mates aren't entirely sure but they've found a pretty good clue: ants never overtake. Not ever. Instead they form into platoons in which all the ants move at the same speed. Increase the density of ant traffic and the platoons simply join together to form larger groups. This is how the velocity remains the same while the density increases. That makes ant traffic significantly different from other types of traffic in which congestion occurs, such as road traffic and internet packet traffic.
That's neat but how have ants hit on this congestion-free solution? Almost certainly through the process of evolution which may well have selected for ants that use the most efficient transport systems.
There's a lesson for traffic planners there, somewhere.
Ref: arxiv.org/abs/0903.2717: Traffic-like Collective Movement of Ants on Trails: Absence of Jammed Phase
Massachusetts Institute of Technology
Comments
Can anyone verify/debunk this? I can't find anything with Google. Either the collective human memory is still better than Google, or else my memory is quite inventive.
dmm
03/19/2009
Posts:207
A small percentage of humans are terrible at estimating distance. In traffic, they will accelerate until the car in front of them fills the windshield and then brake hard until a 'big' gap appears and go on to do it again. This sets up the 'compression waves' seen in one-way traffic.
The MIT folks could do something about this. Make a heads-up windshield display that shows a virtual line across the road in front of the driver corresponding to a 2 second following distance. The drive would then find it easy to place that virtual line under the rear bumper of the car in from for an exact 2- second following interval.
Of course, I know that active cruise control does this automatically and the HUD would just make it possible to do the same thing manually. But, sometimes traffic is easier to manage manually.
bildan
03/19/2009
Posts:20
LOL, you are so right, at least about the behavior! I wonder, though, if the problem is estimating distance. I don't have 3D vision, but do not have a problem following the car in front at a fixed distance. I would more readily attribute this problem to stupidity, or failure to pay attention. If either of these is the true origin, then nothing short of autopilot will solve the problem.
dmm
03/19/2009
Posts:207
I read that small cars are rear ended at a significantly higher rate - presumably by people who judge distance by the size of the car in their windshields.
You've probably noticed people stopping several car lengths from the car in front of them at stoplights and then slowly creeping forward. They simply can't judge distance or where their car will stop so they 'pad' their stopping distance.
These people need all the technical help they can get. We've already given then self parking cars and 'smart' cruise control. I'm sure they'll be first in line to buy robotic, driverless cars - at least I hope so.
bildan
03/19/2009
Posts:20
I was wondering how closely you were following the drivers whose stopping annoyed you.
justahick
03/19/2009
Posts:12
The people I'm talking about couldn't see the car behind them anyway - their mirrors aren't adjusted properly. (If I can't see their face, they can's see mine.)
bildan
03/20/2009
Posts:20
The practice of avoiding, in advance, the necessity to touch the brakes continually teaches us to anticipate speed and distance. A reward in our pocketbooks is always good incentive.
Jim20e
03/20/2009
Posts:2
I've done accident investigations on both aviation and automotive crashes. I can assure you that 99.9% of them are human error as politically incorrect as that sounds.
Humans are always the weak link or as the WWII saying went, "The primary cause of accidents is the loose nut on the controls."
bildan
03/20/2009
Posts:20
covidiu
03/23/2009
Posts:1
jandockx
03/23/2009
Posts:1
2) the ants don't care if they have a few collisions.
dmm
03/19/2009
Posts:207
bsperka
03/19/2009
Posts:2
dmm
03/20/2009
Posts:207
There is a form of depth perception that works at longer distances, but only if a patterned background at a known distance exists (such as a treeline). It does not work against the sky, a uniform background, or a moving background.
Most highway depth perception works with one eye or two, because it is based on the trigonometry of the driver's height above the road and the contact point of the car ahead on the road. But the human analog for angles works on the sine of the angle, not the angle itself. Some drivers do analog trigonometry on the size of the car ahead instead.
MidiMagic
06/10/2009
Posts:5
GaryB
03/20/2009
Posts:71
hdginzo
03/20/2009
Posts:10
Do humans, when walking, create traffic jams?
Secondly, if ants were to drive, would they also create traffic jams?
Walking or running is a great equalizer -- most people (or ants) can walk at about the same speed as each other when walking comfortably.
But people can drive at all sorts of speeds because they are not restrained by physical limitations. Cars allow people to drive at a much broader range of speed regardless of the designated limit (45-80 mph in a 55 mph zone), which means some may be traveling half as fast as others. Living in Houston, I know this to be the case.
This is an interesting article, but stops short of enlightening.
1 -- Should we impose speed limits for drivers mechanically, like they do on Uhaul trucks (they tend to limit the speed to 65mph or so)?
2 -- Should we teach drivers to bunch up?
3 -- Should we institute a no-passing law?
4 -- Should we all walk, like ants, to avoid traffic jams?
ravijsingh
03/23/2009
Posts:1
In the staging area, i measured 200 feet from another jeep, got into mine, and used a grease pencil to mark the jeep's width on the windshield.
On the road, my driver just kept the jeep in front within the marks.
Low-tech, but effective.
Now that I'm out of the Army, I tailgate! :-)
hsfrey
03/23/2009
Posts:13
That simple fact cam explain the "memory" effect in traffic.
mattroche
03/23/2009
Posts:1
Definitely sounds like public transit.
mesagsx
03/25/2009
Posts:1
michael.b
03/25/2009
Posts:1
fbt
03/26/2009
Posts:1
ps. once a review of the old research is done, look at the films. ants do pass, though possibly not the species in this study. ants compose something like 10% of the fauna biomass of the earth. seems that studying more than one species before leaping to assumptions about the entire family would have yielded a more statistically valid conclusion about traffic patterns.
xynthia.delf...
06/10/2009
Posts:1
Automobiles have the following problems ants do not have:
- High mass and inertia (long stopping distance)
- Crash danger of damage, injury, death, and expense
- Crash debris blocking road
- Can't crawl over obstacles or stalled vehicles
- Following distance requirement
- Sight distance obstructions
- Speed limits (ants travel at top speed)
- Intersections (Ant trails don't cross each other's paths - they merge and diverge)
- Breakdowns (Ants don't get flat tires)
- Railroad crossings and drawbridges
- No common purpose (all of the ants have the same goal)
- Drunk drivers and lawbreakers
- Lost drivers
- Load spills to clean up
- Bad pavement
- Cell phones
- Too much hurry
- Time clocks (Ants aren't paid by the hour or fired if they are late)
- Sporting events and concerts that start at certain times
---
If you want to see an ant traffic jam, plug the entrances to the nest.
Another way is to break off the branch they were using, and substitute a new branch they were not using. The ant trails end, they get lost, and they can't pull over and ask for directions.
Drop something yummy near the nest, and simulate a football stadium parking lot jam.
MidiMagic
06/10/2009
Posts:5