Monday, May 10, 2004
Stuff I played with yesterday:
A guy in Seattle's take on behaviour in congested traffic (and how to manipulate it). Via his links page: this interactive model of various congestion inducing scenarios [some glitches, especially with low traffic influx].
Strangely most of the pages seem to cite the M25's variable speed limits - I had no idea it was considered so experimental. I'd always assumed that reducing the speed limit reduces the average speed of vehicles approaching congestion, or in congestion. You reduce the speed, and thereby reduce the shock. Normally people in one lane are trying to do 95[1], whilst there's people elsewhere on the carriageway trying to overtake others doing 55mph. Someone pulls out, someone else has to stand on the brakes, another car sees the brake lights and brakes themselves. And there's a ripple of braking and panic. The people all driving too close to each other all have to brake harder than the car in front [assuming they are unable to exactly match the braking pattern of the car in front, and therefore stay a uniform distance behind throughout. As they can't do this, in order to avoid trying to drive through the car in front, they overcompensate]. They brake harder or for longer, which means they ditch more speed, and everyone behind them has no choice but to slow down to their level.
[1] Yes that's miles per hour, yes people here speed regularly.
The initial shock waves amplify and consolidate until it becomes a permanent wave (which moves against the flow of traffic). The leading edge of the wave continuously envelopes evermore traffic, as the trailing edge leaves vehicles driving slowly away. The traffic leaving the wave has already dropped speed dramatically, and cannot pick it up again as quickly, so forces the wave bound traffic to fall to at least this speed. The vehicles in traffic jams or slow-downs are bound by the surrounding traffic. They cannot escape by going faster as the car in front isn't going faster.
In order to prevent the shockwave initially forming, variable speed limits aim to reduce the speed disparity within the traffic. So dropping the speed limit to 50mph or 60mph [from 70mph], should greatly reduce the proportion of drivers trying to do 95. The proportion doing seventy will also plummet. So now the maximum for most people is about 60, so encountering someone doing 50mph [eg a fully laden mini overtaking a lorry] doesn't cause a massive change in speed. 10mph are dropped, instead of 40 odd (plus the overreaction loss), which means a much smaller proportional change in speed. And because it is a small change it can occur over a longer time, and be much more finely judged. This allows a higher starting speed for the recovery, and can be negated a slight adjustment in the spacing between cars [Car A pulls out from behind a lorry doing 50mph. Car B is travelling in this lane at 60mph. Car B may brake, ease-off, ease-off slightly, or continue as normal. The latter two will erode some of the distance between the two cars, but during this time is likely that Car A is accelerating. If Car B had slowed down to Car A's speed, Car B would now be falling behind Car A. Drivers used to this situation tend not to slow down much, and drift closer to Car A until Car A matches Car B's speed. Once Car A's speed...
Warning - do not find you have to stop blogging for some reason, and leave the post in mid-sentence, as when you come back you'll have a very Kubla Khan[2] moment, and struggle to remember what was the point you were trying to make.
[2] Coleridge was enjoying an [probably] opium high when he started writing the poem. Halfway through he was interrupted, and lost the train of thought, and so the rhythm and content of the poem jumps.
Anyway, point I was trying to make: variable speed limits allow adjustment to compensate for higher traffic volume reducing the capacity of the traffic to absorb shocks. Variable speed limits reduce the variation of speeds within the traffic.
Originally found via Pandamonium. It's probably worthwhile exploring his links, and those of the guy in Seattle (if only for the god-like ability to cause gridlock).
Anyhoo,
A guy in Seattle's take on behaviour in congested traffic (and how to manipulate it). Via his links page: this interactive model of various congestion inducing scenarios [some glitches, especially with low traffic influx].
Strangely most of the pages seem to cite the M25's variable speed limits - I had no idea it was considered so experimental. I'd always assumed that reducing the speed limit reduces the average speed of vehicles approaching congestion, or in congestion. You reduce the speed, and thereby reduce the shock. Normally people in one lane are trying to do 95[1], whilst there's people elsewhere on the carriageway trying to overtake others doing 55mph. Someone pulls out, someone else has to stand on the brakes, another car sees the brake lights and brakes themselves. And there's a ripple of braking and panic. The people all driving too close to each other all have to brake harder than the car in front [assuming they are unable to exactly match the braking pattern of the car in front, and therefore stay a uniform distance behind throughout. As they can't do this, in order to avoid trying to drive through the car in front, they overcompensate]. They brake harder or for longer, which means they ditch more speed, and everyone behind them has no choice but to slow down to their level.
[1] Yes that's miles per hour, yes people here speed regularly.
The initial shock waves amplify and consolidate until it becomes a permanent wave (which moves against the flow of traffic). The leading edge of the wave continuously envelopes evermore traffic, as the trailing edge leaves vehicles driving slowly away. The traffic leaving the wave has already dropped speed dramatically, and cannot pick it up again as quickly, so forces the wave bound traffic to fall to at least this speed. The vehicles in traffic jams or slow-downs are bound by the surrounding traffic. They cannot escape by going faster as the car in front isn't going faster.
In order to prevent the shockwave initially forming, variable speed limits aim to reduce the speed disparity within the traffic. So dropping the speed limit to 50mph or 60mph [from 70mph], should greatly reduce the proportion of drivers trying to do 95. The proportion doing seventy will also plummet. So now the maximum for most people is about 60, so encountering someone doing 50mph [eg a fully laden mini overtaking a lorry] doesn't cause a massive change in speed. 10mph are dropped, instead of 40 odd (plus the overreaction loss), which means a much smaller proportional change in speed. And because it is a small change it can occur over a longer time, and be much more finely judged. This allows a higher starting speed for the recovery, and can be negated a slight adjustment in the spacing between cars [Car A pulls out from behind a lorry doing 50mph. Car B is travelling in this lane at 60mph. Car B may brake, ease-off, ease-off slightly, or continue as normal. The latter two will erode some of the distance between the two cars, but during this time is likely that Car A is accelerating. If Car B had slowed down to Car A's speed, Car B would now be falling behind Car A. Drivers used to this situation tend not to slow down much, and drift closer to Car A until Car A matches Car B's speed. Once Car A's speed...
Warning - do not find you have to stop blogging for some reason, and leave the post in mid-sentence, as when you come back you'll have a very Kubla Khan[2] moment, and struggle to remember what was the point you were trying to make.
[2] Coleridge was enjoying an [probably] opium high when he started writing the poem. Halfway through he was interrupted, and lost the train of thought, and so the rhythm and content of the poem jumps.
Anyway, point I was trying to make: variable speed limits allow adjustment to compensate for higher traffic volume reducing the capacity of the traffic to absorb shocks. Variable speed limits reduce the variation of speeds within the traffic.
Originally found via Pandamonium. It's probably worthwhile exploring his links, and those of the guy in Seattle (if only for the god-like ability to cause gridlock).
Anyhoo,
Oh must find a link, but remember reading a German scientific article that showed that compared traffic to fluid dynamics (or some such) and showed that if a driver in a queue travelling below 80 kmh (50mph) on an Autobahn it would cause a ripple effect ending up in (on average) the 30th car back having to stop completely and hence the stop-so effect for no apparent reason in grid-locked traffic. Below 50mph on the Autobahn counts as gridlocked around here. :-)
Post a Comment
<< Home
