Car2Go + DriverlessCar + Uber - Better than mass transit someday? (Austin: how much, home)

[gpurcell]

Originally Posted by jb9152

Driverless autos will actually reduce either speed or capacity. You can't have both. The only reason that highways are able to achieve the throughput that they achieve today is that we allow human drivers to follow at unsafe stopping distances. Police don't enforce a minimum separation. A driverless car manufacturer, on the other hand, if they're not already buried under a pile of safety regulations, will not take on the liability of building in unsafe operation. As speeds increase, safe stopping distance increases exponentially. Driverless car manufacturers will enforce safe stopping distance, and likely will also include some additional margin to account for less than optimal brake performance, road conditions, etc. So, you have to either reduce speeds, or reduce capacity. Which one do you think people will accept - 70-75 mph on I-35 with long enforced safe stopping, or crawling along at 30 mph with shorter enforced safe stopping but higher capacity?

If we operate these vehicles on highways that have the same capacity constraints that we have today (and are likely to have for a long, long time), then not only will they not replace the need for traditional mass transit, they will actually *reduce* capacity on the system as a whole.

Driverless cars are not a panacea. Yeah, they're cool. But they're not high-capacity, even in huge numbers.

Safe stopping distance is a function of speed of response as well as travel speed. In addition, driverless cars would have the potential to communicate with each other. If anything minimum distances should be shorter and speeds higher.

[jb9152]

Originally Posted by gpurcell

Safe stopping distance is a function of speed of response as well as travel speed.

No, there is an absolute minimum safe stopping distance even at instantaneous reaction times. A catastrophic event up ahead (such as something falling off a truck, a tire blow-out, etc.), if the safe stopping distance is not being enforced (even at reaction time = 0), will result in a crash. Outcome for the manufacturer of the car with the unsafe braking algorithm? Big time lawsuit.

Originally Posted by gpurcell

In addition, driverless cars would have the potential to communicate with each other.

Now you're assuming theoretical improvements (car to car communication) that don't currently exist. It's the old "Step one - steal underwear; step two -; step three - profit!" There's a lot that would have to happen for autonomous cars to "talk" to each other - a lot of infrastructure that doesn't currently exist, at a minimum. Radio spectrum, communication protocols, and likely some type of central back office to monitor and control each car.

Originally Posted by gpurcell

If anything minimum distances should be shorter and speeds higher.

Well, that'd be great. But that's not what's going to happen. Stopping distance increases exponentially with speed, regardless of reaction time. Driverless car manufacturers *will* build a conservative stopping distance algorithm into their vehicles because they are risk-averse. Let one person die because the stopping distance was wrong, and you'll see several things - government swooping in to regulate, and manufacturers protecting themselves by getting even more conservative.

[gpurcell]

You are also ignoring the role of erratic driving behavior in degrading system performance, BTW.

[Novacek]

Originally Posted by jb9152

Driverless autos will actually reduce either speed or capacity. You can't have both. The only reason that highways are able to achieve the throughput that they achieve today is that we allow human drivers to follow at unsafe stopping distances. Police don't enforce a minimum separation. A driverless car manufacturer, on the other hand, if they're not already buried under a pile of safety regulations, will not take on the liability of building in unsafe operation. As speeds increase, safe stopping distance increases exponentially. Driverless car manufacturers will enforce safe stopping distance, and likely will also include some additional margin to account for less than optimal brake performance, road conditions, etc. So, you have to either reduce speeds, or reduce capacity. Which one do you think people will accept - 70-75 mph on I-35 with long enforced safe stopping, or crawling along at 30 mph with shorter enforced safe stopping but higher capacity?

If we operate these vehicles on highways that have the same capacity constraints that we have today (and are likely to have for a long, long time), then not only will they not replace the need for traditional mass transit, they will actually *reduce* capacity on the system as a whole.

Driverless cars are not a panacea. Yeah, they're cool. But they're not high-capacity, even in huge numbers. Since the 800-pound gorilla in Austin transportation is capacity, driverless cars will either have a negligible or negative effect on the single biggest problem we face.

They'll also reduce capacity in that they'll double the number of trips. Instead of a bunch of drivers coming down from Round Rock in the morning, there will be thousands of empty cars heading North _to_ round rock, then back down again. As a commuter with a currently contra-flow, nearly empty road trip, I hate that idea.

Unless they're going to park overnight up there, which means you'll have these 50-100K vehicles sitting around half the time (all night) doing _nothing_.

Edit: I realized this sentence was ambiguous. I meant that each one of these vehicles will probably be in the vicinity of 50-100 thousand dollars (for all the advanced sensors they need, automating the breaking/steering, allocated R&D costs, etc.).

[jb9152]

Originally Posted by gpurcell

You are also ignoring the role of erratic driving behavior in degrading system performance, BTW.

Negated and surpassed by people's penchant to follow at much closer than safe distances at high speed. I drive I-35 between Georgetown and San Antonio a lot. Most of the people I see in the middle and left lane are doing 70 mph+ with a few cars' worth of separation between them.

At 70 mph, instantaneous stopping distance (i.e. no reaction time) is about 230-240 feet, under ideal conditions. Auto manufacturers will add in some distance so that you don't come to a stop right on the bumper of the car ahead (say, 10 feet). They will also build in some margin (maybe, say, 10% to 15%) to account for less than ideal braking conditions. That takes stopping distance into the neighborhood of nearly 300 feet. An average midsize sedan is about 14 feet long. That means the safe stopping distance that will likely be enforced by an automated car is equivalent to about 20 car lengths. Imagine I-35 at 70 mph with every car separated from the car in front of it by 20 car lengths. Think that's a capacity improvement over conditions today, where we allow human drivers to ride others' bumpers at 70 mph+?

[jb9152]

Originally Posted by Novacek

They'll also reduce capacity in that they'll double the number of trips. Instead of a bunch of drivers coming down from Round Rock in the morning, there will be thousands of empty cars heading North _to_ round rock, then back down again. As a commuter with a currently contra-flow, nearly empty road trip, I hate that idea.

Unless they're going to park overnight up there, which means you'll have these 50-100K vehicles sitting around half the time (all night) doing _nothing_.

Good point. How many square miles of downtown would have to become parking lots or structured parking?

[Novacek]

Originally Posted by gpurcell

In addition, driverless cars would have the potential to communicate with each other.

But they can't communicate with all the non-driverless cars. So you're either waiting _another_ 20-30 years for all the non-driverless cars/trucks to be off the roads OR you have to have additional, dedicated, separated lanes for them (gee, sounds like, and sounds as expensive as, mass transit).

[jb9152]

Originally Posted by Novacek

In fact, the driverless technologies are _more_ likely to show up first in, wait for it, rail applications (it's just significantly simpler, a more constrained problem).

They already exist today - WMATA, BART, PATCO, MARTA, LACMTA Green Line, just to name a few that are automated. There is a "train attendant" there to shut the doors and push the "go" button, but the train for sure runs itself.

[jb9152]

Originally Posted by scm53

I was reading something interesting recently (Jarrett Walker?) and the writer was saying we should focus on transit outcomes, and not transit methods. What you are talking about, Steve, is an outcome based solution that is guaranteed to outrage the method focused folks.

Well, since the outcome will be reduced system capacity, I don't think there's much to get outraged over. It just won't work.

[Novacek]

Originally Posted by jb9152

They already exist today - WMATA, BART, PATCO, MARTA, LACMTA Green Line, just to name a few that are automated. There is a "train attendant" there to shut the doors and push the "go" button, but the train for sure runs itself.

Right, but aren't most/all of those cases grade separated? These automation technologies will start to benefit the non-grade separated rail transit. Again, it's a more constrained problem, since it only has to worry about throttle and brake, not steering. And it's easier since it'll probably be going 30-40 mph, not 80 mph highway speeds.

Then it will start to benefit buses (now you'll have to worry about steering, but its easier to amortize the cost over a bus that already costs 500k, plus again, it's not expecting to go highway speeds).

These are exciting and useful technologies, and absolutely worth developing. And they will be useful for single occupant vehicles as well. But it'll be many, many years until it can replace transit. And that's not even considering the efficiency/environment angle. Overhead wire or third rail transit is going to be a lot more energy efficient.