Autonomous Cars

[Greg Locock]

I'd have thought a vehicle was either autonomous or not autonomous.

[Phil]

Driverless trains are the dominant driverless land passenger transport vehicle in the World.  Currently 40% of all passenger trains in cities in Asia have no driver.  They carry100s of Millions of passengers safelyeach day.

To be honest, I don't quite see the issue with the quote "driverless trains are the dominant driverless land passenger vehicle in the world". I'd have probably guessed as much, without seeing concrete numbers. Isn't even the tube in London driverless? In anycase, I did know that Asia has lots of these type of [driverless] trains and we all know that these cities dwarf most western cities in population. It's like saying that mandarin chinese is the most spoken language in the world. Sure it is, but only by population numbers, not by area.

Either way - driverless trains and autonomous cars are very different. First of all, driverless trains are not autonomous - they are controlled and coordinated by humans in a centralized area and they have a strict time-schedule. It's not dynamic.

The challenges of autonomous cars are very different. They are driving in a dynamic and changing landscape with lots of unpredictability.

[Phil]

@henry;

Perhaps my argument is more one of feasibility. I.e. Considering the amount of R&D that is poured into the AV venture, will it lead to a viable business. Will the tech evolve quickly enough to be wanted by enough people wanting to use it?

There are two main challenges I see:

Technologically
The order of magnitude that technology still has to evolve so that a piece of software can drive a car in the majority/all of circumstances without any driver intervention to the point one could say it's as comfortable as taking a cab and trusting your 'pilot' to get you safely from A to B in a timely manner. I.e. an autonomous vehicle that only works in dedicated 'controlled' areas/zones (i.e. highways) may have a more limited appeal than one that can drive anywhere. An AV that requires the driver to be held accountable is also problematic and if this can be solved, depends on how far the governments are willing to go when laws need to be changed to allow the use of fully AV.

Viability
This one is tricky to answer. Obviously, for car manufacturers to build these cars is a great idea. It's a trump card to play and being the first will mean a very lucrative business. The question though I am raising is how much of an investment will governments have to make to allow this to happen? Will we require special roads so that these AVs of the future can navigate the streets safely enough? Limit the environment to be more predictable? I.e. a city with lots of people walking across streets. Could AVs deal with these situations adequately or would we need to separate the two entirely?

Living in a country with a high(er) population density than most (sans Asia), I can already see the challenges to car ownership. Increased running costs, more congestion, longer time travel. Driving a car is losing it's appeal. And it's not because I don't like driving; It's because it's getting too expensive and time travel is increasing due to congestion.

Public transportation has benefits in both efficiency, cost and time travel that cars, regardless of autonomous or not, can't match. Autonomous cars is foremost a project, a multi billion venture by big technological firms (Google) and car manufacturers. Of course they would be, because assuming our population will continue to grow at an exponential rate (with a growing middle class able to afford cars), driving will at some point no longer be sustainable or attractive as a method to get from A to B.

How do you tackle traffic congestion? Build more roads? At the expense of room for residential areas? Build underground? I really don't see any ideal solution here for the future, other than public transportation networks. Assuming that far into the future there won't be any cars at all, one has to ask what these huge car manufacturers will become. IMO - I think this venture into AV is them building for a future where cars may no longer be a very viable business and they may become technological giants for other areas.

Perhaps the difference between my view point and others in this topic, is that while some are looking to what will happen in the next 10 to 20 years, I am looking to what will be in 50 years or more.

[AJI]

...Either way - driverless trains and autonomous cars are very different. First of all, driverless trains are not autonomous - they are controlled and coordinated by humans in a centralized area and they have a strict time-schedule. It's not dynamic.
...

This is the part I really want to understand. Forget the large scale PT v. personal AV argument. It doesn't really matter because both will (may.., possibly...?) coexist.
The question for me is; at what stage does a massively programmed transport system be reclassified as autonomous?
We set the parameters for acceptable behavior. Isn't the system still ostensibly 'us'?

[Phil]

We're probably getting into a bit of semantics here, but following this thread, I'd say "autonomous" to mean without a human controlling it.

An AV vehicle (or the goal of it) is a self-driving car. There is no 'human' in the car in control, nor is there one in a decentralized area coordinating it. It's effectively like those 'robot vacuum cleaners' that roams your house independently and autonomous.

Driverless trains are different. They are coordinated from a decentralized area and monitored closely. If people at a station cause a delay, this needs to factored into the coordination of other trains because you likely have shared train lines at one point or another. Either way, there are people involved who are closely monitoring what is going on. If there is an issue, they intervene, they can remotely control the train etc. This means the train doesn't need any 'AI' or an array of sensors (other than for the purpose of monitoring) to function. It is not a 'robot train' making decisions for itself. It's effectively simply remote-controlled. Not much different than normal trains, with the exception that there's a driver that controls the speed, acceleration, deceleration and can intervene, but the coordination happens outside the train.

Of course there are examples of trains that are nearly autonomous with little to no human involvement. For example shuttle trains that connect two or more terminals at an airport. They probably function in a complete autonomous way. However, these shuttles do not share train lines, they don't need to be coordinated and it is a very controlled environment (a tunnel) with no risk of unpredictability. They don't need any sensors and the 'shuttle' doesn't make any decisions. It's effectively a train that goes from A to B and from B to A and at the station keeps the doors open for a predetermined time until it continues its loop.

[AJI]

Semantics, possibly.., but they (AV's) are still 'programmed' to be compatible with 'human needs', otherwise they're worthless to us, no?
Is that true autonomy?
Even your last example would be programmed based on historical data of human activity, or would it..?

[Greg Locock]

Driverless trains, fine. What about the eight million buses https://www.freedoniagroup.com/industry ... s-3232.htm that move most people on PT? That's what I was thinking of. Plus minibuses.

[henry]

Here’s a announcement of a trial of full size buses in Scotland. [/url]https://www.stagecoachplc.co.uk/media/n ... 1-22a.aspx[/url]

[Greg Locock]

Oh, I know there are autonomous buses, I've been driven in one at an airport. But the point is there are an awful lot of buses in the world used for PT and they all count against this 'dominant' claim.

[Big Tea]

Oh, I know there are autonomous buses, I've been driven in one at an airport. But the point is there are an awful lot of buses in the world used for PT and they all count against this 'dominant' claim.

The thing is that there are just so many ways to 'claim' the stats. Total miles, including where 24 hr operation, passenger miles, miles where the vehicle is automatic, but a driver has to be there, miles on a test track etc, they can dip into a tub and display the produce in a way that fits someones wants.

Aircraft are mostly 'self driving' and if air miles are added it can be huge. As I said in my previous post, I will not disagree with what anyone claims, because there are probably 'stats' to prove it.
Best not to have a falling out with anyone over something so vague.

[AJI]

Oh, I know there are autonomous buses, I've been driven in one at an airport. But the point is there are an awful lot of buses in the world used for PT and they all count against this 'dominant' claim.

Perhaps Dr Currie got a little carried away with the use of the word 'dominant' in his PT good , AV BAD , article, but he did point out that there is a significant amount of autonomous PT out there...

[Greg Locock]

I've just skimmed the paper. I think he was drunk when he wrote it.

[AJI]

I've just skimmed the paper. I think he was drunk when he wrote it.

There's definitely a bit of 'Texas sharpshooter' in him...

[loner]

Ford To Announce 25,000 Job Cuts: Morgan Stanley

the Detroit automaker could soon announce an even larger round of job cuts than rival GM
We estimate a large portion of Ford’s restructuring actions will be focused on Ford Europe, a business we currently value at negative $7 billion," Jonas wrote. "But we also expect a significant restructuring effort in North America, involving significant numbers of both salaried and hourly UAW and CAW workers."
Ford announced last week that it would be slashing shifts at 2 US factories and moving workers elsewhere
Jonas said other automakers will be forced to follow GM’s and Ford’s actions as the industry transforms, first to abandon factories building slow-selling sedans and ultimately to retool to build electric and self-driving vehicles.
https://www.zerohedge.com/news/2018-12- ... tructuring

[Phil]

I still think there is sense in questioning the viability in investing into self-driving cars.

Of course, if you are a car manufacturer and you have lots of expertise in building 4 wheeled cars, you cant suddenly abandon your core business and start building in a different industry entirely.

They are all looking for the same solution to their problem. They want to sell cars, so the hope for them lies in building the next evolution of that (self driving).

I firmly believe self driving cars will be far less desirable by the market in the end. Either they wont work well enough, or there will be legal hurdles to overcome. Lots of these big car manufacturers will go bust (or swallowed up by the bigger ones) because of their huge investments that wont yield the expected turn around.

Either way, the car industry is in a big mess, certainly long term. Environmental and supply problems with fossil fules are forcing them to invest into new technologies (electric driven) and given many are chasing the dream of self driving vehicles, they are forced to invest there as well to not risk being left behind by the competition.

For many, this will likely end badly. For some sooner, some later...