Interesting[0] story from a friend of mine who is a real engineer:
Back where he used to live at some point road authorities increased speed limits and found that average speed decreased. His explanation was simple: once speed limits matched drivers expectations they started to respect them. (Other explanations that I can pull from thin air now is police promising that the new limits will be enforced a lot less leniently than the old, or that speeds went down because of saturation congestion.)
[0]: although unconfirmed for me IIRC, maybe someone here can tell me were to find data
Traffic engineer here. Generally speed limits are set by measuring vehicle speeds for about a week. The 85th percentile speed is set as the speed limit. The reasoning behind this is the assumption that most people will travel at a speed that is reasonable for the facility. Of course, politics often becomes a factor and the speed is changed to fit those pressures.
Other aspects that can be taken into account when setting speed limits is the character of the facility. For example, number of approaches (driveways, other roads, etc.), type of area (urban, rural), and classification of the facility (arterial, collector, local street).
That's the textbook answer. The reality is that most jurisdictions have some fixed limits depending on the class of road. Parks and such, 30kph. Normal streets 50. Highways 80, with 90/100 where there are no ramps/intersections/driveways/people. If we believed in the 85th percentile then we would have lots of different apparently random limits.
Instead we have limits based on polices very much separate from road conditions or actual traffic patterns. School zones are 30kph not because that's the 85th percentile but because that's the law. US highways were once capped at 55mph to save on gas during the fuel crisis (see 1974 Emergency Highway Energy Conservation Act). That certainly had nothing to do with traffic patterns. If 85% was a thing, I wouldn't cross a bridge every day signed for 50kph but where not a single vehicle is below 80 and the cops say nothing until you're over 100.
You are correct, my response is the text book answer. Often the 85th percentile is rounded to the nearest speed that matches the fixed limits. Special speed zones, schools, definitely need to exist for obvious safety reasons.
Yes, I do live in the US. Even here there is a lot of variation on how speed limits are set.
A good example happened near my home recently. A road that connects a small town (population 8,000) to a larger town (population 40,000) was the subject of quite a lot of debate. The road is a rural two lane roadway that sees about 10,000 vehicles per day. Speed studies were performed by three different groups (the small city, the large city, and the State Department of Transportation). Each of the groups determined their own speed limit recommendation, each was different. Ultimately, the State had final say on the matter and split the difference between the three recommendations. The State had originally wanted the highest speed limit of the three. The two towns wanted lower speed limits citing safety concerns and a desire to force alternate routes that would take traffic through areas with better infrastructure for the traffic load.
Very nice, I played a bit around with it but there is 1 thing that bothers me. The sliders all start at the left but the value it represents isn't at the left. Some sliders should start in the middle to be right.
I've seen many of these ring-road simulations over the years. They are great for pitching lower speed limits, but I fault their assumptions. Traffic does not feed on itself in this manner. Aberrant drivers and situations are constantly being inserted and removed from the system. A slowdown at point A does flow up/down to point B like a wave, but it doesn't wrap around back to point A to be amplified again.
I do credit this particular simulation for modeling acceleration/deceleration. Playing around with it leads to the conclusion that many congestion problems can be addressed by increasing the performance numbers for cars, and by encouraging people to use their brakes/engines to their full potential. We should all drive as if in nascar,
minimizing gaps to fully utilize the available road. Clarkson would be proud.
I would like to see a simulation of cars+trucks+motorcyles. Two-wheeled vehicles are common everywhere and in some countries are dominant on the roads, yet they are never included in these models. And toss in a few bicycles just for kicks.
And yet, many believe that the best way to reduce traffic jams is to accelerate as slowly as possible, brake as softly as possible, and leave as large of a gap in front of them as other drivers will allow.
Normal drivers, unaware of the above, will generally also stop and go slowly in traffic jams.
My personal pet peeve is when there is a reason for a slowdown (emergency vehicle, crash, etc) and drivers slow down further to look at it, unaware that they're are then currently causing the jam they just waited in.
I love that it has really good model explanations.
On my simulations though, trucks tended to drive extensively on the left lane. In real life trucks don't usually do this.
This brings back some memories of traffic flow fundamentals.
I do remember an assignment that was modeling multiple different traffic flow models. The thing that stood out to me is how after a long time span with an network that was stuck with the same vehicles, like this model, traffic flow issues tended to become exaggerated.
As a traffic engineer, I use a product called Vissim to do exactly this kind of thing for real world scenarios. Very fun stuff.
It needs a slider for number of 80 year old grandpas who should have been off the road years ago but DMV keeps passing them to be nice.
Our town has implemented a few of these traffic circles. For some reason a memorial to the killed appears a few months each time after one is opened. I don't think it was drivers killed, I think it was pedestrians trying to cross at the inlets/outlets.
Interestingly, this confirms the common-sense wisdom that coasting, rather than braking, helps to even out traffic jams.
If you decrease the "comfortable deceleration" parameter all the way, you significantly increase the number of cars that can be handled without traffic jams.
Also, try turning T/s0/b all the way down, a all the way up, and then increase the density to about 75. Because the simulated drivers have very fast reaction times, the traffic oscillations propagate much faster than the cars are actually moving, creating some cool-looking spiral patterns.
Back where he used to live at some point road authorities increased speed limits and found that average speed decreased. His explanation was simple: once speed limits matched drivers expectations they started to respect them. (Other explanations that I can pull from thin air now is police promising that the new limits will be enforced a lot less leniently than the old, or that speeds went down because of saturation congestion.)
[0]: although unconfirmed for me IIRC, maybe someone here can tell me were to find data