As indicated from the URL, that article is from 6 years ago.
While pooling is a great idea, you run into the congestion pricing problem. Most people use their personal cars at very specific times each day - and pretty much the same times as everyone else. The Bay Area has strong incentives to pool- special lanes - but there is enough spare capacity to open those lanes to people who are willing to pay. There is even more capacity than it appears because of the large number of cheaters.
Car sharing will work great in urban areas where rush hour traffic can be handled by mass transit and sharing can handle much of the rest of the load, like Uber and taxis do today. And I understand many urban residents rent cars when they need them, since that is much cheaper than owning and garaging one.
As for special vehicles for special needs, that is the U-Haul model. And our truck is the pooled vehicle for many of our neighbors who need a big pickup infrequently. We don’t charge them, of course, but it is the same model.
Families are diverse. Do you call a minivan with a booster seat for a 12 year old, a car seat for a newborn, and a carseat for a toddler? Is that particular configuration available right now, on demand? Or am I expected to install them every time I need a ride? What do I do with them at my destination?
Personal ownership (or via personal leasing) isn’t going to go away.
Even companies like Ford are saying that their self-driving cars will be for fleets/sharing initially; we’ll have to wait for prices to come down for personal ownership, and I can’t see that going away in the foreseeable future.
That’s where car pooling comes in. During the middle of rush hour, the roads are crowded with millions of vehicles with an average of 4 seats, each only carrying one person. So if most people voluntarily choose to ride with another person because it reduces the cost the app charges them by 40%, there’s a huge gain right there.
Next, it is not true that all the cars in a city are on the road at the same time. Not remotely true. I don’t know the exact numbers, alas, but lots of people have short commutes, differing work hours, are already at work or are waiting out traffic, and so on and so forth.
If all the cars those people would have taken are deployed during rush hour (since nobody but competing fleet owners actually owns a car), this reduces the total number of cars a city needs even further.
I’d guess the gain would be at least a factor of 4. As in, a city would need, in an equilibrium state, 1/4 the cars and trucks.
If the autonomy ever gets good enough, motorcycle-like gyro vehicles would make sense. These are 2 wheeled aerodynamic shells that balance using internal gyroscopes. Hugely more efficient, and they would also have motorcycle-like acceleration and braking. They’d be extremely nimble - if the computer software were good, it could take advantage of the agility of the vehicle to keep the occupants alive in situations that would have killed the passengers of a conventional vehicle. (because the thing would be able to dodge insanely quickly and squeeze through narrow gaps in traffic if there’s a crash or pileup. These things could also share lanes with each other, riding in tandem to reduce the road space consumed.)
They would be cheaper to manufacture, as well - basically just a shell, a battery 1/3 the size needed for a bigger EV, the front and rear wheels would both be able to steer and have hub motors, a seat, and the usual rack of autonomy equipment.
So this is another way to reduce costs - these things would take only 1 passenger, but be much cheaper, specialized for office worker commuting.
I’ll make a prediction: all this 2020 stuff is crap. You won’t be able to go to your dealer and buy a true self-driving car any earlier than 2024. (I really don’t think it’s going to be until the 2030s, though I’d love to be wrong.)
By ‘true self-driving car,’ I mean pretty much what you’d expect:
[ul]
[li]you tell it where to go, and it goes there in a reasonably efficacious manner (IOW, if it takes 45 minutes for a trip that a human could easily do in 20, that’s not reasonably efficacious)[/li][li]with no human guidance needed between origin and destination, from a passenger or elsewhere (like Waymo’s planned command center where confused AVs can ‘phone home’ and get help from a human operator) [/li][li]obeying all traffic laws (with an allowance for moving with the flow of traffic when traffic is moving above the speed limit), recognizing and heeding all traffic signals, signage, cones, persons directing traffic, etc.[/li][li]Having accidents at lower rates than human drivers in similar circumstances[/li][li]Not being confusing to human drivers substantially over and above the ways that human drivers already confuse each other[/li][/ul]
I’d add ‘not being hackable in potentially lethal ways’ to the list, but that’s one of those only-time-will-tell sorts of things.
I think even 2024 would be great. The Firebug would still be 16 for most of that year, and I expect him to be as flaky a driver as he is in general. I’d pay a pretty hefty premium to start him off in a self-driving car from the get-go.
I’d be shocked if self-driving cars arrived before 2024, and if they do, I’ll cheerfully eat all the crow you care to send my way. I’ll even make it my sig for, say, three months if someone reminds me, something like “I said there was no way we’d have self-driving cars yet. Boy howdy, was I ever wrong.” Some of you folks have long memories. Hold me to it!
Self-driving Chevy Bolts are coming to the streets of Manhattan in 2018
There’s no technical reason that a self driving car would need to communicate with other cars to avoid gridlock. Lots of drones fly themselves with completely onboard computing.
Gridlock happens because selfish and aggressive drivers pull into the intersection when the light is green instead of waiting for there to be space on the other side of the intersection. This can ‘lock up’ an entire city’s traffic grids, because doing this breaks that intersection laterally, and if more assholes do this elsewhere, it can all fail completely.
Other causes are traffic waves, from slow human reaction delays.
Excessive traffic from merges due to blocked lanes, as selfish human drivers pull in front of other cars, forcing them to panic stop.
Humans failing to maintain their vehicles, leading to failure.
And so on. Most of these causes will be prevented if most of the cars are autonomous. The final fix is to make areas with severe gridlock 100% toll roads, with congestion charges for vehicles entering at certain times. Roads are not free, they are a limited resource that cost a very large amount of money and land to build, and people should be charged fees in places where the resource is over-utilized.
Barcelona’s subways have gone from having a driver and “if you feel sick pull the alarm” to not having a driver and “if you feel sick wait till the next station, get off and use the interphone.” The newest trains aren’t so much driven remotely as following the tracks, with the same people who manage track changes pushing the “go” button that makes trains leave the station. The software includes where to brake and stop. Some trains (for example the cremallera to Montjuic) have been autonomous for years: their circuit is just “go up, go down” and they do so at specific intervals.
They’re a lot less complicated than a car but still, it’s a lot less human intervention than just 10 years ago.
[quote=“PastTense, post:66, topic:798499”]
Self-driving Chevy Bolts are coming to the streets of Manhattan in 2018
The article talks about a couple of experimental cars, each with two highly paid engineers inside. And they are limited to staying within Manhatten.
They are not even attempting to produce a fully independent Level-5 car.
This may work as a replacement for the yellow taxi cabs of New York.
For the rest of the world, self driving cars are still decades away.
I know all about car pooling. My dad car pooled in NY from 1952 or so to 1975. He could set it up because he had a rare UN parking pass, because the UN had a lot of people who lived fairly close to each other, and because in those days people worked regular hours. None of this is true in Silicon Valley, at least. I mentioned the car pooling incentives we have. I doubt saving money will be that much of one since people pay $6.00 to save maybe 10 - 15 minutes on a ramp.
The big problem is that you do not want to miss a big meeting because your car pool is there. You also don’t want to spend an extra 20 minutes commuting as your car takes other people home first. Ever hear people complain about other people getting picked up last or dropped off first in an airport van?
Around here rush hour is from maybe 6:30 am to 10:30 and from 3:00 pm to 7:30 pm. Some roads have continuous congestion through the day. Not many people can afford to live near work, alas. And we definitely have very flexible schedules.
Not everywhere is that bad, but good commutes are not going to encourage carpooling, especially if you can relax and rest in an autonomous vehicle.
There was hope that everyone could work from home, but that has kind of petered out.
When you own a private vehicle, you don’t really compute the cost per hour driving since the benefits of having one are so great. If you own a vehicle and intend to make money on it, substantial amounts of down time are going to be a problem, and will force the owner to charge more during congested times to make a profit. You may not be using much gas or electricity from 10 - 3 and from 9 pm - 6 am, but you are still depreciating that car.
Now this is an interesting idea. I think it is less the software, which doesn’t take much space, than the sensors and other hardware which you’d have to fit into this form factor. When autonomous vehicles have been around a while and the hardware shrinks to save costs, very possible.
My grandson’s seat snaps out of a stroller and snaps into a car. Wouldn’t be hard to make them to be easier to install when it is expected to happen frequently. However I think not wanting to make a crying infant wait until the car arrives might be more of a reason for personal ownership.
All depends on the price point - but just like electric cars, there will significant numbers of private ones on the road for those who can afford them, and who consider an hour or two more of work a day well worth it.
Here’s some startup working on a gyro vehicle :
Will “lit motors” succeed and become a brand like Ford or GM? Almost certainly not, but their technology might get bought and used when the time for it is appropriate.
Anyways, I like the concept of these vehicles - you’re reducing the cost to move a car to the absolute minimum. An aerodynamic shell like that uses the least energy to move at high speed. Two wheels means better performance and less tire wear. The vehicle is short and narrow, so it eats less road space. It needs a smaller battery. Smaller motors. You could make the front and rear wheels the same, where both can steer, both have hub motors, both have a backup mechanical brake.
You’re correct that you’d still need just as many cameras, still need lidar and radar and all the rest.
What do you consider an hour or two more of work a day? You mean, working during the commute? If so, I’ll consider that work time, starting during my commute. And I fully expect to be able to afford to be an early adopter (I get a company discount, after all).
In my particular case (common, but definitely in the minority), I would expect that I’d actually be more productive, because I spend a lot of time visiting suppliers and plants. Instead of concentrating on the road during that one hour trip to Auburn Hills (or six hour trip home from Louisville I just finished!), I could do something useful like catch up on my 115 unread emails.
Hell, for this added productivity alone, I should think my company would give me a car!
I hope you noticed that the article was from 2012 and predicted production availability by 2014. Now it’s three years later and it doesn’t look any closer to being available.
The rest of my post that you quoted :
Will “lit motors” succeed and become a brand like Ford or GM? Almost certainly not, but their technology might get bought and used when the time for it is appropriate.
The fact that some startup ran out of money is hardly proof their idea is bad : to make an automobile you can sell costs billions of dollars. This idea seems viable.
Exactly this. There are plenty of execs around here not rich enough to have a chauffeur but making medium six figure incomes. They drive themselves, and are on the phone all the time. shudder Even though in California it has to be hands free. They will consider being able to read reports and do email well worth the price of the car.
I lived only 17 miles from work and my commute was almost 2 hours on a good day, and the road I took isn’t even the worst.
Where I live, the traffic is quite bad, so I don’t even attempt to drive when the roads are not available for full speed travel. At one job, I would leave at 5:15 am and get to work at 6am, then go home at 3pm. At another job, I would leave at 9am, get to work at 10, and work til 8 to 9 pm.
Certainly is harder to do this in some ways, but it beats blowing hours of your life in traffic.
A car like that would be extremely dangerous in a collision with a regular car, or God forbid a truck. And even with a fixed barrier. If you want a car to be as safe as a normal car, the crumple zones are going to have to be about as big. No matter how strongly the car is built, its dimensions dictate the distance over which the passenger compartment is going to decelerate in a crash. The shorter the distance, the quicker the deceleration, the higher the G forces on the driver. If you want a safe single-seat car, it won’t be all that much smaller than a regular compact car.
I’m very skeptical of the idea of driverless cars happening any time soon. Everyone on Reddit and elsewhere seems blindly optimistic about it in the same way they’re all convinced that we’re going to colonize Mars soon just because Elon Musk said so, irrespective of the actual engineering challenges. They seem gleeful at the idea of manual cars being banned and all travel coming under the control of whichever corporation happens to provide the software.
This has been mentioned before, possibly in this thread, but the possibility of mass hacking of driverless cars is terrifying. Someone hacks into millions of cars remotely ala Stuxnet, or an insider-either a crazy person or an agent of a foreign power, who works for the car company-plants obfuscated malicious code into a software update. At a certain date and time, millions of cars simultaneously crash at high speed into the nearest solid object, then stay on full throttle, overheating the tires, batteries and motors until they catch on fire and burn any survivors to death. Millions dead, larger than a small nuclear war. The economy is wrecked because everyone is afraid to get in their car and go to work the next morning. How can you stop that from happening? Or, if not done maliciously, then by some bizarre software bug that causes the same sort of thing.
Plus I’m extremely skeptical that computers will have the intelligence to deal with all the unusual situations that can crop up. Are things like image-recognition so reliable that we can stake lives on them? There’s an object overhanging the road. Is it a thin tree branch that you can drive right through, or a steel beam that will decapitate the driver? You’re parking in a field. Is that brown area up ahead simply a flat patch of dirt, or a shallow puddle, or is it a deep pond of muddy water that will drown all of the passengers? How does it deal with road construction, conflicting road markings, hand signals from a cop?
From what I read, the first Tesla self-driving death came from a car that drove straight under a semi trailer because its radar was working in two dimensions only, with no vertical dimension, and so it perceived the trailer as two separate cars with a gap in between. That’s extremely primitive in comparison to the sort of object recognition capabilities you would need for a safe, true self-driving car.
If people’s travel comes under the control of software, that raises all sorts of concerns about freedom and privacy. Will it report the driver’s travel habits to the company and/or government? Will the government be able to disable a car because the owner is wanted, or to enforce a curfew due to civil unrest? Will the software company, being a private entity, be allowed to revoke someone’s self-driving software license and refuse to do business with them because the driver posted something politically offensive on Twitter? (Thus rendering them unable to work, or to buy groceries, because remember, manual cars will banned as “too dangerous”.) These are all highly disturbing scenarios.
I wouldn’t limit this just to execs, but engineers, sales reps, project managers, pretty much anyone conducting professional business and are at mid-career or higher with supporting income.