Robocars As the Best Way Geeks Can Save the Planet 394
Brad Templeton writes "I (whom you may know as EFF Chairman, founder of early
dot-com Clari.Net and rec.humor.funny)
have just released a new series of futurist
essays on the amazing future
of robot cars, coming to us thanks to the DARPA Grand Challenges.
The computer driver is just the beginning — the
essays detail how robocars can
enable the cheap electric car, save millions of lives and trillions
of dollars, and are the most compelling thing
computer geeks can work on to save the planet. Because robocars can refuel, park
and deliver themselves, and not simply be chauffeurs, they end up changing
not just cars but cities, industries, energy, and — by removing
dependence on foreign oil — even wars. I, for one, welcome our new robot overlords."
(More below.)
Templeton continues: "The key realization is that while the safety and timesavings that
come from having computers as chauffeurs is very important and can
save a million lives every year, a number of interesting consequences
come from the ability of robocars to drive themselves while vacant.
This allows them to deliver themselves to us on demand, to park
themselves and to refuel/recharge themselves. On-demand delivery makes
car sharing pleasant and allows the use of "the right vehicle for
the trip" on most trips. Self-refueling means the people using
cars no longer need care about range or how common fueling stations
are, enabling all sorts of novel energy systems with minimal "chicken
and egg" problems. Because passengers don't care about the range
of their taxis, battery weight and cost are no longer issues in
electric cars and scooters."
Wow, good job! (Score:5, Interesting)
Robocars can only exist after lawyers are killed (Score:5, Interesting)
The reason cars don't drive themselves is not a problem of technology, but of liability. Now, if there is an accident the driver is blamed. Carmakers are unwilling to take on that liability and themselves be blamed for accidents.
Re:Public transportation (Score:5, Interesting)
Go deeper into the article about the end of transit. Buses are actually quite inefficient, because while loaded at rush hour, on average they carry few passengers. In the USA, city buses use more fuel per passenger-mile than cars do -- on average. And none of the other forms are a great deal better, though some do beat cars. Lightweight electric vehicles are 10 times more efficient than buses. It's one of the key realizations about transit in the article.
KISS (Score:3, Interesting)
Why not start by coupling a frontal sonar and the gas-brake control to enforce the safety distance? Easy to do, and could save a lot of lives.
Re:Wow, good job! (Score:5, Interesting)
I hope I don't gloss over this (read the roadblocks section.) There are many technical and political problems to solve.
As for energy, the goal is to use far less energy than we use today (whether it's cheap or clean is nice but orthogonal) and it's far from limitless.
The AI is not so powerful. Most animals can navigate in traffic of their own kind, even insects. But no, it's no tiny project -- but it's a tractable large project.
You don't need to solve traveling salesman! In fact, I believe centralized control is a bad idea. You can solve traveling salesman over small problem sets, it's only trying to solve it for large numbers that's explosively NP.
You just have to do better than we do today.
An excellent web site (Score:5, Interesting)
I give a talk on the consequences of Moore's Law to a freshman class every year, and one of my topics is autonomous vehicles. This web site does a great job of summarizing the future of personal transportation. A few other points I discuss with the class:
(1) Mass transit as we think of it will largely vanish within 20 years. Cities will find it far easier to maintain fleets of robocars, and dispatch them right to the doors of residents, rather than maintain traditional subway and bus lines. The "last half-mile" problem of getting from the door of your home to the door of your destination will be solved.
(2) The authors discuss "sleeper cars", but they don't really consider all the ramifications. A huge chunk of overnight business travel (everything within a few hundred miles) will be taken over by robocars. People will go to bed in the sleeper car, open the door the next day, and find themselves at their destination. Consequently, hotels and motels will offer short-term rooms (for one or two hours) so that people can shower and dress on the road. A significant portion of the U.S. population will literally become nomadic, sleeping in robotic RVs every night, and waking up to a new destination every morning.
(3) Once robocars are widely accepted, human drivers will be forced off the roads very quickly. How? By 100% enforcement of all traffic laws with high-tech imaging (also thanks to Moore's Law). A human will be unable to conform to the ultra-rigid driving laws that robocars will handle with ease.
As I say to my students: "You are the last generation that will need to learn to drive. To your children, it will be an option. To your grandchildren, knowing how to drive a car will be as quaint a concept as knowing how to saddle and ride a horse."
Re:Sounds like... (Score:4, Interesting)
While there are certainly advantages to living in geographically self-contained units, there are also massive benefits to centralizing industries.
Yes, the "slow foods" movement will tell us, accurately, that shipping our produce from hundreds of miles away causes an incredible amount of waste in fuel.
But consider the alternative - a small farm for every nine city blocks. Suddenly, instead of having a system where one farmer can produce food for a thousand, you have a system where one farmer produces food for, say, fifty. Which means you have to have 20 times more farmers. Which means there are fewer people to provide other services. The same goes for commerce: five corner stores might be more convenient than one larger, more centralized 7-11 - but now you have five times as many people working in low-end retail.
It's centralization of the more menial services that allow so many of us to have jobs in less immediately-necessary services - like programming or science. And almost-completely-unnecessary services, like video game design and filmmaking? Forget about it.
Also check out Marshall Brain's "Robotic Nation" (Score:3, Interesting)
Re:Wow, good job! (Score:3, Interesting)
a city that has a large cab pool is the best place to deploy this technology, you're right, robot cars don't eliminate the need for a car, but cabs are very useful in new york city, for say someone who rides the train to work, then wants/needs to take a cab because they didn't drive a car into the city.
and who says robot car driving systems won't come with personal cars? if they can do it realistically for cabs, then they can do it for personal cars. also, a robot cab lets you redesign cabs entirely, you no longer need a to use up '2' spaces for a cab driver, and an empty chair, or just an empty mounting bracket for a chair... suddenly you can have 4 person or 6 person cabs, without making special cars for them.
Re:Wow, good job! (Score:4, Interesting)
Re:Wow, good job! (Score:3, Interesting)
Re:Wow, good job! (Score:3, Interesting)
They won't be deployed until we are convinced they kill far fewer people than humans do at the wheel. In fact, I suspect in the USA a very low problem rate will be demanded. I think the engineers can deliver that rate.
I could be wrong on that. It's worth trying.
The number of people killed by human drivers in the USA is similar to that killed by Alzhiemer's or stomach cancer or several other major killers. Largest killer from 5 to 45 in fact.
If I told you, "We could probably cure Alzheimers with a relatively modest engineering effort" would you ever write a post wondering if it was worth it?
Not on roads as we know them. (Score:3, Interesting)
I am in full support of this vision. However (and unfortunately), I think the practical answer will resemble robotic trains more than robotic cars operating on the current network of roads. Plus, the main benefits of an improved transportation system will involve restructuring the way cities and communities are built when they are not sliced apart and divided by acres of roadways.
First of all, while there has been some limited success in building autonomous cars, but we can't even get autonomous airplanes accepted into our air transportation system even though planes have practically been able to fly themselves for decades. Hell, most cities can't even get people to accept conductor-less subway trains, and have to hire college students or bums to sit in the front cabin.
The robotic vehicle would have to be completely isolated and separated from unpredictable human traffic and other sources of interference, if only for liability issues.
The best first step in widespread use of robotic cars might just be on the interstate highway system, where they could construct a special lane designed only for robotic vehicles. So you could drive your car/truck onto an interstate, auto-merge into the robotic lane, set the autopilot for your destination exit, and take a nap or otherwise entertain yourself until an alarm wakes you up to exit.
For incursions into urban areas, you'd want something similar to the Personal Rapid Transit (PRT) systems everyone was investigating in the 70's. Take a look at the CabinTaxi system at: http://faculty.washington.edu/jbs/itrans/cabintaxi%20photos.htm [washington.edu] . There are modern PRT systems finally being planned for deployment recently in Heathrow and Dubai... however, they seem to be limited to airport shuttles and aren't really large enough to meet the promise of a large distributed network with many stations.
Speaking of Dubai, the biggest obstacle will be financial, of course. The road and highway system is expensive, but a lot of the infrastructure is paid for by the user in purchase and maintenance of their own personal vehicles. While the city as a whole would find the entire system cheaper if the government would purchase and maintain a smaller number of shared vehicles, good luck convincing them to finance both the network and the vehicles if they can just build the network and have the users pay for their own vehicles. Of course, car sharing companies such as Flexcar / Zipcar offer something of a shared vehicle, they only have limited potential unless they'd allow one-way rentals... where you can pick up a Zipcar at one "station" and drop it off at another "station", where someone else could make use of it. You'd need some way of getting the cars back to empty stations, but that would realize benefits in terms of reducing the area of pavement needed for parking if everyone had their own personal vehicle.
However, I don't think advanced transportation is the magic bullet that will solve all of our problems... I think much greater benefits will be realized by redesigning cities to be denser, more human friendly, and carfree (check out http://carfree.com/ [carfree.com] ), so people simply don't need to travel so far from a nice home to a nice place to work.
So yes, I'm an Arcology nut (check out my MSSE thesis on my homepage). I think the Dantzig / Saaty "Compact Cities" book from 1971 had the most comprehensive plan for constructing a city that I have seen in my research (you'll have to look it up in a good library, it's fairly rare).
In any case, I agree that this kind of development should be a national priority, since there is a *lot* of room for improvement. But since improving the place you live and how you get around are kinda mundane, "infrastructure" issues, I figure we'll see little to no advances in the Western world until China develops the technology and discipline and manages to dust us with their production efficiency, and maybe eventually a high standard of living (said only half-jokingly).
Re:Wow, good job! (Score:4, Interesting)
Why do we even need AI for this? There's a wonderful invention called the railroad that has moved people and goods for nearly 2 centuries. Railroads take less energy. No steering required, all that is needed is throttle control.
Whether or not rails are used, AI can be dispensed with by putting dumb signal readers in the cars, and having senders all along the roads. I recall reading of just such a project some years ago. We have signs all over the place advising drivers how fast to take corners and other such notices. Shouldn't be too difficult to make that all automatic. The vehicle will do only 3 things: don't collide with other vehicles, move at the speeds the road tells it to, and enter and exit the road system when the passengers wish.
Re:Only in the US (Score:3, Interesting)
I'm not sure I'm following their logic but they give examples. Sorry, I don't have the reference for the article with me.