Robocars - Autonomous Drive, Self-driving, Driver-less cars

But they are not arguing to delay the role out because of this are they? they are just saying it needs discussion and thought put into it. Gandalf is correct (as I pointed out 9 days ago ). Doesnt mean that they should not consider difficult questions prior to legalisation of AVs.

Simplified much.

Very much so.

Show me the source code EdZ...

Send me a mail offline.

As with every other invention, the legal aspects lag behind. Look at internet vs. privacy laws, encryption vs. snoopers charter for recent hot topics.

These vehicles already have been allowed on the roads, before any of these ethical, legal, insurance aspects have been considered, certainly in any detail.

If you try and stick a bunch of laws in, like you're getting at, the the tech could stall and we all lose out. Sensible states in the US (and the UK) are letting car designers/developers get on with it.

So no, they should not consider difficult questions prior to the legalisation of AVs. They already haven't.

If you try and stick a bunch of laws in, like you're getting at, the the tech could stall and we all lose out.

I'm not. this is what I said 9 days ago

All this legal stuff around who's accountable, insurance etc. (and possible the updating of laws to fit the new model) could potentially delay autonomous cars by years. I wonder how many deaths will occur in that time. The moral thing to do would be to role out what we have now and see what happens.....

But they are not arguing to delay the role out because of this are they?

I have never said that one should "delay" anything only that one should not run into things without throught.

they are just saying it needs discussion and thought put into it.

It is more than that. We need to look back and learn from other distruptive technolgies what happened. Recall that the movement for constructing highways in the US was started by a cyclist lobby. We need to see how the transition from horse to automobile happened--- and see how corporate power also influenced the development of towns, cities and transport. Los Angeles, for example, had a metropolitan transport system with trams and a planned metro in 1908:

In 1926 elevated trains were put to the vote--- e.g. Chicago-- but lost.

I think get it out there asap and start work on the difficult legal stuff asap, including potential moral dilemmas. I would just go with

i.e. we'd likely be fine with an instruction that if a collision is inevitable then the car should minimise kinetic energy transfer to all participants, and leave it at that.

until/if a better solution comes up. My worry would be that when the gen pop starts to use AVs, innovative companies trying to push development will get sued out of existence using current legislation due to accidents not of there making. This could stall the tech.

I guess a lot of what will happen will be press hysteria, irrational moral outrage and right-wing columnist frothing.
We'll save thousands of lives a year, in a flash, once full scale roll-out occurs. But there will be accidents, and there will still be lives lost. All it would take is one kid to run out in the road and get killed and the incandescent rage would blame it on the robocar. It'll take a while for such stories to go away - in fact, accidents might be so rare that any time they did happen, you'd still get a big news story.
Anyway. If I Robot taught us anything, it's that Shia LeBouef shouldn't be allowed in films.

I'm guessing that this mesh-network with heart rate monitors will not be implemented in the first wave of technology.

You don't need heart monitors.
http://arxiv.org/pdf/1509.05016v1.pdf

Abstract— Anticipating the future actions of a human is
a widely studied problem in robotics that requires spatiotemporal
reasoning. In this work we propose a deep learning
approach for anticipation in sensory-rich robotics applications.
We introduce a sensory-fusion architecture which jointly learns
to anticipate and fuse information from multiple sensory
streams. Our architecture consists of Recurrent Neural Networks
(RNNs) that use Long Short-Term Memory (LSTM)
units to capture long temporal dependencies. We train our
architecture in a sequence-to-sequence prediction manner, and
it explicitly learns to predict the future given only a partial
temporal context. We further introduce a novel loss layer for
anticipation which prevents over-fitting and encourages early
anticipation. We use our architecture to anticipate driving
maneuvers several seconds before they happen on a natural
driving data set of 1180 miles. The context for maneuver
anticipation comes from multiple sensors installed on the
vehicle. Our approach shows significant improvement over
the state-of-the-art in maneuver anticipation by increasing the
precision from 77.4% to 90.5% and recall from 71.2% to 87.4%.

Then why did you say this just now?

Doesnt mean that they should not consider difficult questions prior to legalisation of AVs.

Didn't the automobile companies buy up the shares of and then remove the public transport from US cities?

Didn't say they need to slove things prior to legalisation of avs. You seem to be reading what you think I am saying, not what I am saying.

Maybe I misread an earlier post. I dunno but you seem to be contradictory. I'll go back to the Ale thread where more sense is to be found.

Didn't the automobile companies buy up the shares of and then remove the public transport from US cities?

The Huntingon company only sold their system in 1945... and yes it was purchased by a consortium of, among others, Standard Oil of California, Firestone Tire Co and General Motors. GM actually got caught and paid a $5000 fine. But well before that these companies set their sights on buying up the metro trains they did what they could to sell busses and promote private car ownership. To this aim they started bus companies and "did what it took" to see that they got the contract to operate.. it was a hard sell but already by the mid-1930s GM was deeply entrenched in politics. Their success was limited since the trollys were owned by the public electricity utilities and so they could not purchase any and the quality of service of the trollys was good. This changed in 1935 when the Federal government forced the utilities to sell their lines by 1938. GM, of course, happily purchased the trollys and "updated" them to "more flexible" buses. The irony is that "Public Utility Holding Company Act of 1935" was intended to regulate electric utilites and protect consumers against "big business".
In reminding us of this act we too see where "good intentioned" legislation intending to protect consumers resulted in something else-- a result that could have been forseen.

you are clearly missing the points...

Edward, my very good man, there is a saying that I like to use on occasions such as this... "Bullshit baffles brains," and you spout a lot of shite.

There is only one point. Can the idea of autonomous vehicles be 'sold' to the public? The debate on the technology is a moot point, we already know that within 'x' time Google or another will have a vehicle that is a better driver than 99% of drivers on the road today. So the question remains, will a global fleet of autonomous vehicles serve more people in a safer, more efficient and cheaper manner than the current solution? If the answer is yes then that is the path to take.

Legislation comes with the benefit of hindsight, as with all laws.

Hypothesize > invent > create > release > iterate ad infinitum until we have a system that is as good as it can be under whatever the 'current' situations.

Now they're coming for the pavements... https://thestack.com/iot/2015/11/03/self-driving-delivery-robots-starship-skype-london-in-2016/

There is only one point. Can the idea of autonomous vehicles be 'sold' to the public?

It won't, I think, be "sold" as a commodity to the public but as a service. The various pillars needed to provide this service are already falling into place. Munich Re: for example
http://www.munichre.com/us/property-casualty/publications-expertise/global-topics-experts/autonomous-vehicles/index.html

The debate on the technology is a moot point, we already know that within 'x' time Google or another will have a vehicle that is a better driver than 99% of drivers on the road today.

Not really that hard. We are effectively, from a purely functional perspective, with the technology almost there... that is, however, not the whole game changer. There are still a number of pieces to put together.

So the question remains, will a global fleet of autonomous vehicles serve more people in a safer, more efficient and cheaper manner than the current solution? If the answer is yes then that is the path to take.

That is the same argument that could have been given in the 1930s for buses. The implications and the total cost to society-- of polution, global conflict etc. --- did not enter in those balance sheets. The impact on society was also not really considered. The promised liberation and clean air afforded by the motorcar, I argue, became enslavement, smog and a social "nightmare"

Is there a dytopic side to the self-driving "utopia"? In a society driven by uncontrolled capitalism I could imagine not just cartels but a push to ban human power and force a winner takes all approach to mobility.

Shall it be a good thing when motorcycles and bicycles are removed from the public transit space?

Legislation comes with the benefit of hindsight, as with all laws.

Good legislation needs to come from vision and with foresight.

Hypothesize > invent > create > release > iterate ad infinitum until we have a system that is as good as it can be under whatever the 'current' situations.

What is good? Reminds me of the role of chemists in WW-I with Fritz Haber and Victor Grignard advancing chemical warfare as the more humane of weapons.

I have no idea what you are talking about and I have a strong suspicion you don't either!

Is there a dytopic side to the self-driving "utopia"?

I don't know, but if there is, it would make an interesting sci-fi novel. Go!

Now they're coming for the pavements... thestack.com/iot/2015/11/03/self­-driving-delivery-robots-starship-skype-­london-in-2016/

More here: http://mashable.com/2015/11/02/delivery-robots/#SWotv1qPIsqL

How soon after they're released will they be taken apart by people like, err, tech people and programmed to fetch, ooh let's see now, beer?

And how much will they be available for

And how much will they be available for

I suspect that they won't be "for sale" but as part of a service.

How soon after they're released will they be taken apart by people like, err, tech people and programmed to fetch, ooh let's see now, beer?

I don't think it will be any mystery to what is in the machines nor how they'll work. Whatever "mystery" shall be in the data used to train the networks--- and that, just as with the other players, shall remain trade-secret. Building the hardware for little robot vehicles just as building drones is right now about as mysterious as building bicycles. It should be pretty quick order to determine what sensors they use...Lidar, radar, IR, etc. as well as what GPU boards they are probably using.. If I was them I'd probably use an NVIDIA board.. probably a Jetson TK1 or maybe something X1 based.. there are a few other companies making K1 and X1 boards and it is not difficult to make ones own custom boards but I would not yet bother.. they are cheap. Heck.. maybe even Drive-PX (that is what Tesla get) if I could cut the right deal with them...

I was inferring that they'll be stolen and/or vandalised before they can complete their first delivery.

And then I was asking how much he wants for a stolen one.