A company called SkyTran is making maglev passenger “cars”
secondnexus.com
secondnexus.com
Something like skyTran is going to be super important for future cities. A regular gasoline-engine car can go 150 km/h, but in our cities they spend half of their time stopped, and so it's common to get 15 km/h in real life. And then you have to find a place to park it! Traditional massive subways and elevated trains are better, but building them is too expensive to space them closely — so unless you're lucky, you have to spend 20 minutes or more walking to and from them. And then they still have to stop every 400 meters or so, because you're sharing the car with other people. So you still end up averaging something like 15 km/h.
A personal rapid transit system like skyTran can have lines spaced very closely — as closely as streets, or more closely — and can actually deliver close to its peak speed of 100 km/h. You could get from any place in San Francisco to any other place in under ten minutes, instead of the hour-plus it often takes at present.
How important is that? The entire benefit of a city is that you're close to many people doing many things. It's being close to people that makes us put up with the pollution, noise, and undesired social interaction that come with cities. But with our current city transport systems, you're within 20 minutes of only about 80 square kilometers, about half a million people. A system like skyTran effectively brings you much closer: you would have 3500 square kilometers within that same distance, about 25 million people, if density didn't change.
That is, with skyTran or similar personal rapid transit systems, you could live in a place that looks like San Mateo or Redwood City, except without the cars — but enjoy even more convenient access to city amenities like good jobs, artisanal toast, live music, and offbeat special interest groups like Postgres, than you can get today by living in the heart of San Francisco.
There's a critical mass needed to get benefits like this, and I suspect it's close to the half-million people I mentioned above: on the order of a couple hundred kilometers of transport line, densely covering an area of several tens of square kilometers: an up-front investment on the order of a billion dollars. But maybe skyTran's Tel Aviv project will surprise me and transform Tel Aviv more than I expect.
In https://news.ycombinator.com/item?id=9788092 I did the calculations to show that you could get substantial benefits by replacing even the elevators in a single skyscraper with a skyTran-like system. The numbers only get better as the system gets bigger. Scaling, far from being a problem, is the solution. Eventually you will run into bisection bandwidth problems, but only at scales far, far larger than our current cities. A SkyTran track transports pairs of people at about 3Hz; 200 parallel SkyTran tracks across the width of San Francisco could transport SF's entire population from one side of the city to the other in 22 minutes (plus the transit time of 6 minutes), or 11 minutes if they shared cars.
I'd love to have these instead of cars though. Then I can finally buy a bicycle.
The difficult thing about scaling something like SkyTran will be transferring between tracks and merging at high-speed.
But even if all of that works nicely, you're still limited by the time people take to get in and out of their cars. You'll end up with traffic jams waiting for 20 people to get out of their cars before your car can park at the station...
The actual number SkyTran is designed for is 3.2 cars per second, but I rounded down for ease of calculation. That gives you 9 meters between the fronts of cars, not 5. It turns out that when you can decelerate at human-body-safe 6 gees, which you can do on a SkyTran track but not a regular railroad track or street, then you can stop from 100 km/h in 6.6 meters.
As I've explained above, although conventional trains do work on long, straight stretches, they do not in fact work very well for short-range trips, delivering end-user speeds an order of magnitude lower than their peak speeds. Even that they can only deliver at the cost of enormously expensive infrastructure to support the massive cars needed to make conventional trains cost-effective.
Long, straight stretches are not necessary for SkyTran; you can go just as fast around curves as you can on straightaways, because the cars are suspended from the tracks. Some other PRT systems do not have this advantage. You need to keep acceleration loads manageable, but that's not difficult.
Merging at high speed is a trivial problem with a centralized real-time control system. If you can ensure that cars on the same track are at given relative positions to each other, then you can just as easily ensure that for cars on different tracks that are going to merge to gether.
I'm not familiar enough with SkyTran's design for transferring between tracks to know what problems you see with it. Can you elaborate?
If you had everyone going to (or from) the same station at the same time, you could have station traffic jams. (And I think SkyTran's per-car capacity of 2 passengers is probably suboptimal for this reason, among others.) But since the stations are all on sidings, that only happens if you're going to the same station and there isn't a siding free at that station. You can solve this problem by going to another station nearby — I think the station spacing for most places would ideally be around 50 meters, but if that isn't enough for high-traffic areas, you could build stations there every 25, 10, or even 5 meters, thus further shortening the walk in the average case while removing the bandwidth bottleneck in the overload case.
This makes PRT a fine idea, by itself, at medium densities, but troublesome in the context of existing road and rail systems that are set to be fully roboticized and "shared" in coming years. Each time we squeeze more capacity out of the existing paradigms, PRT loses a little bit of its appeal.
The silver lining in this cloud is that the general idea of coordinating traffic better is very promising. Several new PRT-style ideas have a notion of a "modular" highway that smoothly loads taxi-style shared vehicles into trains while in motion, eliminating a major transfer delay. Because the vehicles are shared you don't get the parking issues or the stops that dog traditional rail - just a very smooth, fast point-to-point with a bit of climbing in and out of cabs.
The biggest weakness of all existing transportation is that it's "lumpy" and oscillates between overcapacity and overcrowding very quickly, whether it's a highway or a subway. What changes is how this turns into a user disadvantage - either in vehicle congestion and heavy land use impact, or frequent stops and crowded cars.
Here is a good guide for anyone interested in winter cycling:
[PDF] https://faridrener.files.wordpress.com/2011/09/winter-riding...
[1] http://www.cbc.ca/news/canada/winter-cycling-tips-1.3314804
Maglev trains in general seem like great, cost efficient technology. They are generally opted against because they require a different type of rails. Given all the other high-speed train options, building a second rail network is nearly never worth it.
This leaves only inner-city transportation to maglev, where there are no shared rail tracks, and where maglev can shine with near-silent operation.
I don't really see how smaller train cars change any of this. You could build one-person train cars on regular train tracks. You would have to upgrade all the collision-prevention systems used by regular trains, but that's already being worked on to allow better track utilization.