This isn't going to work for the most obvious case: rush hour. Which is the case that is the root cause of most issues.
If you're still stuck with peak road usage nearly equivalent to that today, your goal isn't to solve prominent issues today.
If you're still stuck with peak road usage nearly equivalent to that today, your goal isn't to solve prominent issues today.
This would require collaborative networked/distributed self-driving, which is not the same as the let's-use-this-as-an-excuse-for-AI-research individual self-driving we have today.
But really most people shouldn't be commuting anyway. WFH should be much more of a thing, even if it's not full-time.
Meanwhile, we have solutions which work today, several of which can be done today. WFH, incentivizing working outside peak hours, building more densely and closer to cities, investing in public transport, and more. We just don't want to do it.
We see this in The Netherlands. Public transport has gotten noticeably worse, car usage is going up as a result, and roads are expanding to compensate. In a country where housing is a massive problem, which means people will move to less desirable places (read: places further away from work hubs). Now we have a chicken-and-egg problem with regards to public transport, and increased car usage is pressuring space which could be used to create more homes and remove cars from the peak.
We simply don't need to wait another 10-15 years for self-driving to finally be a thing. What needs to be done, is accepting that things will suck for a bit to then get better eventually. Continuing on the same path with self-driving cars will only stall the problem instead of solve it, anyway.
I’ll quote from one of my favorite books Algorithms to Live By:
“It’s true that self-driving cars should reduce the number of road accidents and may be able to drive more closely together, both of which would speed up traffic. But from a congestion standpoint, the fact that the price of anarchy is only 4/3 as congested as perfect coordination means that perfectly coordinated commutes will only be 3/4 as congested as they are now.”
So you are assuming there are essentially no more humans driving?
Also, if you want to improve capacity, how about bicycles and buses?
> Even with failures, likely less dangerous than leaving critical decisions to individual human drivers.
You know what's even less dangerous, like essentially no danger, train.
> This would require collaborative networked/distributed self-driving
So the most complex possible solution that is 100% unproven and even in the best cases is far worse then having a city optimized for walking, biking and trains?
Like I just don't understand. Why do you start with the most inefficient solution possible, and then try to apply (expensive) technology to try to make it better.
How about you start with the most efficient, cheapest technology and apply that in 60% of the cases. Then solve the next 35% of the problem with existing technologies that already solve these problems.
And then for the last 5% you can try to solve them with some amazing future tech.
Its quite simple, design cities to be walkable. Make that safe and a priority. Then extend that by the most energy efficient (and space efficient) mode of transport, bicycles. Then use trains to connect different walkable parts of the city with each other.
Then at the very last step, maybe have some fancy self driving cars for a few special cases.
We know this works. It has been done. And its not expensive, it in fact safes money.
Whether it is self-driving with an incredibly optimized algorithm (good luck with that) or manual carpooling, the same problems are still going to bubble up. The solutions already exist. We, as a society, just don't want to deal with the consequences.
Carpooling failed because it doesn't make sense to use amateur drivers on the same overfilled road network to carry a couple more people. If the entrance to the city is backed up, car-pool lanes don't add much.