SoylentNews is people
SoylentNews
The New York Times has a story about what may be a more likely future of public transportation.
A small electric bus chugged along at a slow but steady seven miles per hour when a white van, entering the street from the side, cut in front of it. The bus slowed, as if its driver had hit the brakes, and got back up to speed after the van moved out of the way.
But this bus has no brake or accelerator pedal. It has no steering wheel, either. In fact, it doesn't have a driver — it operates using sensors and software, although for now, a person is stationed on board ready to hit a red "stop" button in an emergency.
At a time when self-driving cars are beginning to make progress — most notably with a trial program that the ride service Uber began in Pittsburgh this fall — the bus represents a different approach to technologically advanced transportation.
I say a more likely future because of the following:
A driverless car, after all, is still a car, carrying at best a few people. By transporting many passengers on what could be very flexible routes, driverless buses could help reduce the number of cars clogging city streets.
Few advantages accrue from driverless cars if the streets and highways are clogged with them. The passenger(s) can curse the vehicle up ahead instead of its idiot driver. My take: The idea has some promise, especially in places where people do not have long distances to travel.
(Score: 2) by maxwell demon on Monday November 07 2016, @07:36PM
Why do you think the system would be more expensive than the human driver?
The Tao of math: The numbers you can count are not the real numbers.
(Score: 1) by khallow on Monday November 07 2016, @10:24PM
Why do you think the system would be more expensive than the human driver?
1) Lot of gear for very few vehicles. We're not speaking of just the automated gear on each vehicle, but also the command and control gear for the whole system. And it would be modding vehicles costing up to a million dollars apiece.
2)Challenging routes.
2) Everyone who touches the automatic driving system inherits liability, including the municipality. That's going to be an expensive cluster even in a European court IMHO.
3) What happens when the system goes down? Either you don't run critical routes or you have emergency drivers who probably will be full time employees drive those routes. There will probably be other inefficiencies in the system at the beginning too.
If you can get the costs (including hidden ones like liability) to under a few million dollars per bus, it might work out. For example, if your effective labor costs of drivers are $25 per hour and you want these buses running for 6000 hours a year without any drivers, then that's a labor cost of $150,000 per year to displace or $1.5 million over ten years. If say you're a moderate sized city with 100 50 passenger buses, the system would need to cost you less than $150 million over ten years. A big problem here is that one really big bus accident could blow that budget (and it might not even need to be one of your own buses! A jump in insurance from accidents in related systems might do it alone).
(Score: 2) by Grishnakh on Monday November 07 2016, @10:45PM
I don't know where you get the idea that the equipment is expensive. George Hotz has been working on a system that fits on certain Honda models and is based on a Raspberry Pi IIRC. Many modern cars all have the necessary hardware built-in: forward-facing radar, electric steering, electronic throttle and brake control, etc. All you have to do is add some cameras, and small cameras these days are dirt cheap (just look at the cameras in cellphones). The most challenging part about building a driverless car now is the software, mainly because it has to be absolutely correct.
The "command and control gear for the whole system" is a simple server on the internet. That's not expensive, and doesn't require anything too beefy if you're talking about a "moderate sized city with 100 50 passenger buses", unless you have some team of absolute morons building the thing and bloating it beyond belief (which is indeed a valid concern these days, considering how much horrible and bloated software is out there).
BTW, big-city bus drivers cost a LOT more than $25 per hour in total fully-loaded costs. Try more like $50-100/hour.
(Score: 1) by khallow on Tuesday November 08 2016, @03:54AM
George Hotz has been working on a system that fits on certain Honda models and is based on a Raspberry Pi IIRC. Many modern cars all have the necessary hardware built-in: forward-facing radar, electric steering, electronic throttle and brake control, etc. All you have to do is add some cameras, and small cameras these days are dirt cheap (just look at the cameras in cellphones).
You could do that, and it might even be capable of driving on a street, but it won't be legal.
BTW, big-city bus drivers cost a LOT more than $25 per hour in total fully-loaded costs. Try more like $50-100/hour.
As I noted, there's room to cut some costs before we go with automated driving. Doesn't require any engineering either.