[lotsofpulp]

> We'll be finally able to traffic jam free when autopilot will make us slow appropriately in high density conditions, avoiding accordion-like brake/accelerate/brake cycles

No, because each vehicle has different acceleration and deceleration rates. Watch any road with an incline and you will see congestion develop relatively quickly before the incline.

And a traffic jam is inevitable when the bandwidth of a road has been exceeded. Each vehicle needs a certain amount of space in front of it, and behind it. The faster it goes, the more space it is, as well as if it’s heavy. This means for a given number of lanes, at a given speed, there exists a max number of vehicles. Once this maximum is exceeded, speed must drop as it is no longer safe to travel so close to vehicles in front and behind, and so bandwidth also drops. Hence, traffic jam. The different accel/decel rates contribute to the congestion also, as vehicles need varying amounts of time to fill the space in front of them.

Bottom line, there is no solution to traffic jams other than reducing space needed to travel Per person (trains instead of cars), or increasing speeds (not possible in cars on current roads without exceeding acceptable safety risks), increasing the number of lanes (space is usually not available), and the easiest option, reduce the number of vehicles traveling the road via variable tolling based on congestion (currently the best option in my opinion).

[Retric]

Humans don’t closely approach the theoretical limits of highway capacity. They encounter a much lower limit and get stuck in a range of suboptimal choices.

[lotsofpulp]

I can see it helping a little in cases where congestion is caused by some less than ideal braking, but in urban areas where the congestion is frequently due to merges or in a hilly area where trucks are going slower in one or two lanes, I don’t see it helping much. In rush hour, the cars are bumper to bumper, it’s because of all the cars entering the same roadway needing to merge in.

Also, I would assume any congestion alleviated would be offset by induced demand by the people who now see less congestion and decide its worth it to jump on the road.

[dsfyu404ed]

>I would assume any congestion alleviated would be offset by induced demand by the people who now see less congestion and decide its worth it to jump on the road.

Proportional to the number of people forgoing trips because the congestion sucks. You'll never eliminate all congestion at peak hours but you can make "peak hours" into "peak hour" if you make things more efficient. A more efficient system can also withstand more traffic before things start backing up into each other and it all goes to hell with feedback loops.

Scenarios where individuals are planning their own movement (road traffic, foot traffic in a stadium, etc) can generally be modeled like a plumbing system carrying a highly viscous and compressible material. If you throw every trick in the book at it to make all the "features" (intersections, doorways) efficient you can reap a lot of benefits without actually widening bottlenecks.

[lotsofpulp]

Yes, but I suspect the main bottleneck is an individual taking up 12+ sq m of space in the pipe.

Attacking the other parameters of the problem is nice, but not the best use of resources. But perhaps that’s all we can wish for at this point.

[HeWhoLurksLate]

Breaking news:

Everyone in the USA is required to drive a 6 sqm hybrid Fiat, effective immediately.