Take control when necessary
Given the spate of hacking stories (e.g. http://www.theregister.co.uk/2015/07/24/car_hacking_analysis/) then pedals & wheel may not be required for driver (or someone) to take over anyway!
Ford has claimed that it will be mass-producing a self-driving car by 2021. The Michigan-based multinational’s president and CEO, Mark Fields, has informed a Palo Alto audience that the business would be selling cars without steering wheels or pedals in five years' time. This, according to the Financial Times, is because Ford …
"That's going to be a barrel of fun if there's 40+ ton of truck 20 feet behind you at 55.99 mph at the time...."
55.99mph is 82.1 feet per second, which if that's your relative speed to the truck, you have 1/4th of a second to react, take control, & do something to avoid a collision.
Which ain't gonna happen.
I think that the big motor companies are just waking up to a new business model.
For years now, people tend not to "buy" a car, but rather get it on some sort of leasing deal, which allows them to roll over to a new model and just keep paying. This is actually more profitable for the auto manufacturers than the traditional purchase route, and it offers them the chance to provide additional services.
I still think that we will move to some sort of subscription based package; you'll pay based upon what type of vehicle you want to use, when you tend to use it, the locations, and how long you want to wait for a car. Those that don't want to wait pay more; if you can wait half an hour, you'll pay less.
It will probably be linked to some sort of mobile technology, either through phone app or similar. The prospective user will request the vehicle, and it will be dispatched either from a holding location, or after having dropped off a previous customer. Having completed the journey, it will then either park or go to another job.
The key advantages; fewer cars on the road, more efficient use of those that are. They will also get maintained and valeted on a set cycle, so should be kept in better condition. The motor companies will probably do deals with 3rd party maintenance companies, so there is the opportunity for smaller businesses to get involved.
I still think my £50 bet is looking OK!
I share your vision though I think it might take some time arriving. The insurance companies are going to hate it and fight every step of the way as this will cause a serious contraction in insurance revenues. The big motor companies will probably end up providing very very cheap insurance for their autonomous cars which are effectively all being driven by the same driver with a very well understood risk profile. Traditional insurers will be left fighting for the high risk human drivers where losses will be both common and expensive. So premiums will rise pushing more and more people into autonomous cars leaving an ever diminishing market for the insurers.
A very acute post. I must admit that the potential consequences for the insurance companies had not occurred to me. With essentially only "one driver" to insure for the whole fleet, insurance costs would indeed drop dramatically and the auto-manufacturers could very well afford to supply insurance as part of the deal. See icon.
"this will cause a serious contraction in insurance revenues."
I already see the answer to that coming, and I don't like it.
Mandatory traveler's insurance.
Insurance companies will lobby for laws that anyone getting into any vehicle, car, bus, aircraft, ship, whatever, must have fully paid up traveler's insurance for that type of vehicle, or no travel for you!
Most likely a default "personal" policy allowing travel in land based vehicles.
A "business" policy that allows air travel.
A "Holiday" policy for air/ship travel.
Etc, etc, and so forth.
The Future is coming... for your wallet. :/
"I think that the big motor companies are just waking up to a new business model."
In the 1980s, the USA Ford division hardly made a cent from building cars; Ford dealers and FoMoCo made money from financial agreements with "owners" of new Ford cars. It is not a new business model.
It may be a new way of identifying suckers.
I don't see it meaning fewer cars on the road. People not owning their own cars may mean less cars in total, but more of them actually on the road. Not having your own car parked and waiting at the start of your journey will mean a shared car having to drive to you, for however long that takes that's an extra car on the road (rather than parked).
Taxis have taxi drivers.
If autonomy can be achieved at relatively low cost, the key changes of eliminating drivers and reducing collisions* would allow vehicles to be moved optimally at significantly lower cost. That favors taxis and jitneys over scheduled services, ride-sharing and pooling over solo commuting, and hire over ownership.
Really huge potential shifts.
* Not accidents
According to the US Department of Transportation, the average age of cars on American roads is 11.4 years. That fact alone tells you that despite all of the marketing hype, American drivers don't run out an buy (or lease) new cars just because they contain new technology.
All this hype reminds me of the Tesla CTO claiming, in 2010, that by 2015 that 20% of cars on the road will be electric. His prediction missed the mark by a factor of 100. The percentage of cars on the road that are electric is 0.2%. Had he bothered to do any due diligence on the average age of cars on the road were in 2010, he'd have known his prediction was impossible to meet.
"the average age of cars on American roads is 11.4 years."
I bet the average taxi on the road is much younger then 11.4 years and that is the market Ford is initially targeting. Once you eliminate the cost of drivers from taxi services it is likely that relying on taxis for transport will become cheaper and more convenient than owning your own car - at least in cities.
"ou'll pay based upon what type of vehicle you want to use, when you tend to use it, the locations, and how long you want to wait for a car. Those that don't want to wait pay more; if you can wait half an hour, you'll pay less."
The only time I thought this had a real shot at working was the home/office/home commute. It's the most common journey and I could see you could justify a 1 person vehicle tightly tailored to your personal environment IE cup holder, decent radio, jacket hanger etc.
I figured they'd be delivered by 1 driver in a train that would drop off the end vehicle and self park ready for its driver of the day. After the driver parks it it's then collected, washed and cleaned (or used by shoppers to come from their shopping) before dropping it (or one just like it) back in time for the drive home (and recollection).
But the rest of the time I don't think people want something that anonymous. They want their vehicle, their way, in their colors, parked where they left it.
for tears of laughter streaming down the face?
And into which market will they be selling these cars exactly? The one with the legislation that allows steering-wheel-less autonomous vehicles? An unproven, untrusted technology? On the mass-market? It's somewhat easier to develop new technologies that are compatible with existing legislation, such as hybrids, rather than try to force legislation to keep up with technology.
OK, I'll bite. I would buy one. I enjoy driving, but the overwhelming majority of it is a chore. Road deaths are at unacceptable levels. It's governments job to legislate for the benefit of the population. I'm not going to feel sorry for them that they have to work a bit harder/faster to make the necessary changes.
You'll buy one too. You might not want to, but once the insurance premium of a normal car skyrockets above what you can afford (it won't take long) that will be that.
Sorry, I know this sounded confrontational. Road deaths aren't funny though.
Road deaths are at unacceptable levels.
The UK has one of the lowest rates of road death in the UK, so logically you'd expect other countries with higher rates to be more keen on self-driving cars, yet they aren't.
The (possibly unpalatable) fact is that the level of read deaths is acceptable to most people. Clearly not to the friends and family of any specific fatality, but at a general, non-specific, level we're all OK with getting in our cars and venturing out on the roads in what we feel is reasonable safety.
If a government wants to ban something on the grounds that it saves lives, then banning smoking would have a larger impact with fewer economic effects. In terms of what to spend taxpayers money on, cancer kills 40x as many people in the UK each year as car accidents do. Car accidents are way down the list of causes of death in general, and really don't merit the level of hysteria they often trigger.
"but once the insurance premium of a normal car skyrockets above what you can afford (it won't take long) that will be that."
Why would the insurance premium skyrocket? There is no monopoly, the premium is largely based on the risk and free market. For the premiums to skyrocket the risk has to skyrocket, If introducing self driving cars on the roads increases the risk of collision, noone will do it in the first place. Nor the prices of cars will skyrocket, so the risk stays roughly the same, hence the insurance premiums.
"If introducing self driving cars on the roads increases the risk of collision, noone will do it in the first place."
BUT, if driverless cars prove themselves to be much safer than meatbag controlled cars, the premiums on meatbag controlled models will increase, as they will be proven to be a higher risk.
Your logic is false. If the driver less cars prove much safer record their premiums will be lower than the meatbag models indeed. But that doesn't mean the meatbag premiums will get more expensive than meatbag premiums now, only than driverless cars in the future.
Not to mention that it will take years if not decades to correctly establish driverless cars safety records to match a driver with say 20 years of experience and 5+ years of NC bonus, So I wouldn't hold my breath for this to be the selling point of driverless cars
Any insurance company worth its salt will be making sure it covers potential outlay from driverless collision incidents. With barely any data, I expect they will take a far more cautionary approach unless it is mandated otherwise. e.g. based on the engine capacity etc of an equivalent vehicle. That's just collision damage, of course. Then you've got loss through theft and fire and other means (e.g. the idiot box drives through a swollen river calculating that it's just a ford). Or the manufacturers will be forced to underwrite the vehicle, which kind of makes sense; they should put their money where their mouth is.
Insurance premiums are not going to increase for human-driven cars. 1) A lot of your insurance premium is used to pay for the risk to you from all the crazies on the road with you in their sights, so if driverless cars REALLY are safer, then your insurance premium will probably go down as a result. 2) Driverless cars don't vote, but drivers do, and in most places in the U.S. the state insurance commissioner is an elected position. If someone starts jacking up insurance rates to artificially push driverless cars, I project a great political future for politicians willing to take the contrarian position.
Well said - who cannot see the problems with this idea immediately?
The idea of self driving cars being on the same roads as the boyracers who drive like idiots round the area where I live is just ludicrous.
"Where we're going, there are no steering wheels" - my guess is hovering somewhere between the scrapheap and the courtroom. What's your corporate plan for the first time a kid runs out in front of one of these and gets killed? More technologically, how does the current state of the art in software development remotely match the horrendously complex task of driving safely through an urban environment? Autopilots in planes are working in a vastly simpler place.
This is good marketing hype for the general public, it can't seriously hope to work in a busy city.
Autonomy has two easily definable challenges:
1) Drive on the road to your destination
2) Don't hit stuff you shouldn't on the way
Given the advances in machine learning, processing and video, I don't see it as anywhere near impossible.
If a child runs out in front of an autonomous car, my corporate plan would be to check the logs and recordings and determine whether the car reacted appropriately. Clear evidence showing it reacting faster than any human could will likely help the case a lot.
Governments _will_ want autonomous vehicles. Too many benefits. If the tech works, it'll show up in stats _really_ quickly.
The fitting of seat-belt mounting points became mandatory in 1968, and cars built between 1965 and 1968 had to have them retrofitted. It was 1983 before it became mandatory to wear them in the front, and 1989 before back seat passengers had to join in. Despite the vast amount of data collected worldwide in the 50s and 60s, it still took the best part of two decades and God knows how many attempts to pass the legislation before it happened. And that's for a basic safety device, rather than anything game changing like steering-wheel-less vehicles with all the liability and insurance issues. So I don't think that any mass-market is going to open up for them within 5 years purely because of the speed with which legislature moves. Possibly there's a market for self-driving taxis in a controlled environment in some progressive car-free towns or business parks, but by 2021 I can guarantee there's not going to be the kind of volume of sales that matches their announcement. I can see a main production model with advanced guidance technology as an optional package, perhaps that's what they mean, and if they'd have said "we will be making this technology available as an option on some of our main production models, operating in all modes as high as local legislation permits" I'd not be laughing, I'd be nodding and saying, yes, I can see that happening.
Someone once told me that since my car was so old I should take out the rear seat belts so they didn't need to be used.
I told them I fitted them, they did not understand why I would fit them.
Perhaps I didn't want to be hit by a low flying passenger?
Really confused the late night Police in the quiet lanes, pull over old hatch pretending they were in the RAC rally and find a couple of well strapped in drunks.
Pretty sure that they stopped me having a full checkover more than once, not that it would have been an issue except for the exhaust and running a larger engine than on the documentation.
In the UK? IIRC about 1/3 of all US drivers demand to assert their constitutional rights by not wearing them.
Which is why US airbags are loaded with more gas generating powder.
The explosive force of which has killed at least 1 7 YO child who was in the front passenger seat of a car that tripped by mistake. It's not good if you're a small women either.
So I guess self drive cars are quite attractive in America.
Or will this drive Ford into another bankruptcy and necessitate yet another USG bail out?
I have discovered that at least one country, Finland, does not require a human driver. I'm not sure about their equivalent of the Construction and Use of Road Vehicles Act, where many mandatory features are related to the driver (such as the field of view, window shading, mirrors etc etc). The UK has a special prototype vehicle legislation to deal with all these circumstances where a vehicle inherently cannot comply with the existing legislation for whatever reason.
Ones with fewer regulations and fewer potential multi-billion dollar lawsuits than the US and UK. How about China, or India? Or maybe getting into the rental/taxi market for tourists in third world countries?
OK, I'm skeptical too about this technology being ready to the point where you can get away with a car without a steering wheel by 2021. How will you steer it if you live out in the country and want to drive it into your backyard to unload some stuff? Using a smartphone app? Yeah, I'm sure no one will hack that...
"How will you steer it if you live out in the country and want to drive it into your backyard to unload some stuff? Using a smartphone app? Yeah, I'm sure no one will hack that...
Not to mention roads which are not on maps, roads which ARE on maps but not actually there, road improvements where a junction/exit is closed, but there's a temporary one. eg, for those that know the area, Scotch Corner to Dishforth being upgraded 6 lane motorway on a live, running road, new junctions, changed junctions, closed junction and best of all, I'd love to see what an AI driven car can make of the northbound exit being closed at Scotch Corner, the temporary exit being half a mile before the currently closed one. Yes, a truly "smart" car will report the changes back to base, but someone has to be first.
Apart from anything else, the last time Ford decided they could tell the buyer what they wanted rather than the other way around was such a roaring success that it's still a staple case study in marketing text books.
I look forward to being able to call my self-driving Edsel to come and pick me up from the pub.
There was that recent crash where the automation missed seeing a white lorry turning across it. If it can't see that, there's all sorts of situations that could confuse it.
Some years back I was behind a bus that had a black and white photo-realistic advert painted all over the back. So lots of different greys in different blobs. I missed seeing that bus until I was very close to it. They had actually managed to camouflage a London Transport double-decker bus.
And then there's stealth cyclists. Not the usual ones who ride at night with no lights, but the ones who also wear dull colour clothing and remove the reflectors and anything else that might make them detectable. I came very very near to hitting one of those once on an unlighted road where I'd dipped my lights for oncoming traffic.
How are automated cars going to deal with those if they can miss a white lorry?
"How are automated cars going to deal with those if they can miss a white lorry?"
Because for every widely-publicised incident like that, there are 100 non-newsworthy incidents of preventable collisions being caused by human drivers which would have been avoided by autonomous vehicles. You so eloquently described how you were misled in a similar situation; how do you even know that your average driver might not have crashed into that white lorry too?
C'mon, did you see pictures of that truck? Absolutely NO WAY a human would have crashed into it! The horizon excuse was bullshit, it was 3PM.
You're right that at some point (I don't think Tesla has nearly the data to prove it yet, Musk knows the times when "autopilot" is activated are not times when humans are mostly likely to get into fatal accidents) autonomous cars will have fewer accidents/fatalities than human driven cars, even when driven by a human with years of driving experience who is following the traffic laws and paying full attention. Obviously then for the good of society as a whole it makes sense to encourage use of autonomous vehicles.
The problem will be things like that truck collision - where the autonomous car gets into an accident that a human never would. They will never be perfect, and the type of stuff that can fool machine vision / sensors is different from the type of stuff that can fool/confuse a human driver. When the autonomous car kills people in other cars or pedestrians in a situation where a human likely wouldn't, it will garner a lot of publicity, and the Luddites and those who want to hang onto the ability to drive their own cars will come out in force.
Laws will be required to limit liability for such incidents, currently it is essentially limited by the fact that most people carry $1 million or less of liability insurance, and most don't have even that much in personal assets to go after. If you can sue Ford Motor Company on the other hand, the lawyers get dollar signs in their eyes...
Given the current state of development, true driverless cars (no manual control option) are a long way off. They might be ok for town/cities and motorways etc, but elsewhere?
A driverless car would never cope with the back roads and lanes where I live (Somerset/Devon border). Most of the roads are like dirt tracks with Farmer Giles on his tractor weaving around for added fun. Not to mention free ranging sheep, cattle and horses on some parts!!
Nothing wrong with having a "hands free" option, but manual control would be a must.
With current technology, the car can't tell whether the blob in front is a sleeping sheep or a newly fallen rock. It's rather difficult for a human being too. On one hand, a driverless car is less likely to be distracted by shagging ponies. But human eyes are better at identifying sheep from rocks than AI.
Google's driverless cars are not designed for real llife. They are designed for suburban commuter trips by middle class G people. Today, driverless cars are permitted to perform a limited function on public roads; top speed is limited so driverless cars have to take the back routes; driverless cars are not being tested on freeways, highways or motorways -- there is always a flesh and blood backup.
Driverless cars are designed to solve problems that middle class kids perceive. Will a driverless car turn up for a disabled kid living on a back lane in rural Dorset? Nah, we don't have the maps for that. But we know a faster route from Islington to St Pancras than any cabby with "the knowledge", assuming there is no congestion. Assuming.
My favourite challenge: the 1907 Peking to Paris race in reverse (not reverse gear) for driverless cars. France to the Ukraine is a cruise -- after which there are few maps and the driverless vehicles get into pickles. Driverless cars don't solve any problems on the two thirds of this planet which is land: there aren't any roads.
Driverless cars sort of fix a problem in the developed world -- take me home from the pub if I am pissed. That's a Google solution for G people. Where sheep don't sleep on the road.
Read a similar story to this on Ars Technica that gives a bit more detail...
The cars will effectively be driverless Uber type cars for around cities, not for general sale to the public.
They will also have more sensors such as lidar and very detailed mapping.
The story on the Beeb includes the quote
"This is really a car designed for very specific urban environments. It’s a car that’s going to take people at 20-30 mph through city centres."
I don't know what the link to the Financial Times explains as they want money before they let me read the article.
Why do people keep bringing up the "Tesla death crash" story? Autopilot is only fancy cruise control, which is Level 2 autonomy that requires the driver to monitor the car at all times. Ford are aiming at Level 4 and AC wants Level 5 .
I agree. Everyone has fixated on a car crash involving a system that was never meant to completely control the car and bring up "what if" scenarios about white trucks and cyclists, while conveniently ignoring the number of white trucks and cyclists that get hit by regular cars with humans at the controls every day.
Granted the Tesla autopilot was poorly named and was being misused by the guy who got killed. But you're ignoring the fact that autonomous cars have to be FAR better than human drivers to gain public acceptance. BTW I do NOT mean the average human driver whose accident numbers are mostly due to those who are texting/drunk/distracted/inexperienced/old/speeding/tired, but human drivers with a lot of experience who are following all traffic laws and paying full attention to driving!
Any accidents at all they get into that a good human driver paying full attention likely wouldn't have that result in death or serious injury (particularly if it was a pedestrian or cyclist) is going to be a huge PR issue. And this WILL happen, because the sorts of things that can fool or confuse software & sensors are different from the sorts of things that can fool and confuse the human eyes and brain. They will need to have data to counter it, showing how autonomous cars are significantly better drivers than the top 1% of human drivers.
"It’s a car that’s going to take people at 20-30 mph through city centres."
Not in many capital cities it's not! it will be lucky to manage10MPH, especially if the things really take off. Think of thousands of Autonomous cars going into a city centre with safety a high priority in their programming, they will go slower not faster, even if the software and machinery has the ability to go at speed the Elfin safety crowd will make sure that a dangerous automatic machine will be tethered in some way.
Autonomy as I have said before, is not the way it will go if they take our steering wheels it will be a central computer running transport and where's the fun in that? ( Particularly after your autonomous Fusion goes critical with a self aware central mainframe at the figurative wheel.)
Be honest, how many hours does your car sit doing nowt, on the drive at home, in the car park at work.
I accept that individual requirements dictate our personal need.
But if I could just call up a car for the 4 or 5 trips I have to take each week, it could potentially save me a fortune. I guess it depends on how much those 4-5 trips cost me I guess.
Now if I want fun, I drag out the bike, can't see them being autonomous any time soon.
My car sits in my garage at night and in the company carpark by day but I don't want to share it with anybody else.
It has my sails in the boot if I want to go sailing in the afternoon, some tools in case something needs fixing, chocolate biwcuits on thepassenger seat, my music on the stereo and most importantly the seat and mirrors are adjusted for me.
I could probably save a fortune right now by using taxis but I'd rather spend the extra on my own space.
Taking your comment to its logical conclusion, I don't need my house when I'm at work nor my wife when I'm sleeping, do you believe that people will share those as well ?
Well at least the important bit is covered, you can already get cars that adjust the seat and mirrors depending on the driver (by recognising different keys) - not that you need mirrors cos its self driving and you also won;'t need to adjust the seat. Music - ohh, if only there was an IT solution to that - maybe the music could be stored on some electronic thingambiginny? The choclate biscuits probably melt and go stale anyway, so we are down to sails & tools.
I think a more valid gripe about a 'shared' car is that there won't be one around at rush hour, when they are all busy, or after 10 pm when they are all busy recharging themselves,
If a car is parked, then it's not on the road. It's not burning energy. If a self-driving taxi is wandering around empty, then that needs more road and more energy. Taxis and self-driving car equivalents require MORE road space and MORE energy than a similar size vehicle that only moves when productively occupied.
Everyone just assumes the opposite. Which is daft.
When pressed on this daft assumption, the hand waving about future magic efficiencies somehow more than compensating for the now-newly-obvious-to-them direct negative effects of cars wandering around empty.
The big saving will come with ride sharing.
Once the majority are hiring a ride rather than owning a car the scope for a master system organising shared travel can take off - a cross between mass transport and a universal taxi service.
I imagine a system of different sized vehicles for different uses. Mini bus sized pods perhaps powered by super capacitors for local hops, larger/smaller vehicles for group and individual longer journeys.
Yet they're only just now doubling their research facility workforce to a whopping 260? This from a company that is yet to offer a five year warranty on it's cars and only gives 12mths on everything on the inside?
Yeah, okay. When they're released I'll send my robot valet out to buy one. On a hoverboard.
I'm glad Ford announced they are working on this; however, I'm not exactly holding my breath.
I remember seeing a Ford commercial 15 or so years ago where they talked about having active radar in combination with HUD (heads up display) information for enhanced night driving vision. Looked cool but still haven't seen a car with that stuff on it.
I don't see Ford or GM pulling this off. They're idiots. Complete morons. They couldn't make a round wheel work if it wasn't legacy technology by now.
This is like NASA announcing every year that we're going to Mars in 15 years. That's been happening for at least the last 50 years I know of, and we're obviously still not on Mars.
Ford and GM know that they aren't delivering an autonomous road car in the near future. They know that it isn't going to happen.
Google -- a firm which hasn't built a series production car-- continually yacks that it has a car for the open road -- the Google open road. How fast do the Google cars go? Does it work on a freeway?
The solution would be in the take control mode to have a popup joystick or app become available for your phone that would allow control AFTER you enter your drivers license number so it can be verified by phone call with the subscription service, or you have to provide your dl number to have the subscription. Insurance still carried by the company though. Maybe a call to emrgency services if things really go wrong Onstar style. think thrustmaster flight stick on a hidden panel. Everybody knows how to use that these days.
Had a discussion just this morning, gist of which was: who, exactly, is pushing for autonomous cars?
More specifically, who expects to profit?
Somehow it keeps coming up: insurance companies, not so much insuring autonomous owners, but positively soaking the drivers of non autonomous cars for at least a couple of decades under the guise of "increased risk."
Aside from accident avoidance, and some utility on long highway trips, I'm not convinced there's a market for these unless drivers are forced to buy them.
Aftermarket chip reprogramming already is an automotive thing, and there will be a brisk market for disabling these things.
being an old fart who passed their driving test in the mid 80s I am really looking forward to self driving cars.
This side of not having a job like Jezza Clarkson and his mates have still got mucking around on a test track, driving to me is just a chore.
Plus with self driving cars a lack of automotive mobility in dotage is not something I'll have to worry about. (I'll probably have other mobility problems, primarily caused by sitting down in front of a computer too much in my middle age )
The standard of driving I saw on a recent tour of England just leads me to believe self driving cars can only be a good thing. Admittedly most other drivers were pretty good, but there is always a small minority of idiots who really should just volunteer themselves to be crash test dummies at Thatcham.
Its now looking like well get self driving cars before self driving trains. Which is just plain unbelievable, considering how much easier it is to make self driving trains.
By now, you should have read the story about the Roomba (robot vacuum cleaner) mindlessly spreading dog poop all around somebody's house. If not, look it up. Well worth your time.
The lesson is duh-obvious. It applies DIRECTLY to all self-driving cars, and should be recognized by the over-confident idiots that naïvely believe they can conquer the 'A.I. is hard' barrier.
Will the self-driving car have ears (microphones) and software to hear the fire truck's siren? Will it hear the terrified passenger screaming, pleading, begging the damn car to stop due to some impossible to predict and complicated circumstance? Can the car recognize terror in a human voice screaming in an unknown language? Any human can. Anyone think of this sort of requirement? Human drivers are not permitted to wear headphones while driving, but self-driving cars can all be deaf?
Will the car have a nose, to detect the smell of itself being on fire? The dragging brake. The mattress wrapped around the driveshaft. Can it correlate the local news, and distinguish various smells? Can it see the brush fire on the next hill, and realize that the burning smell isn't car-like?
Will the car have vibration sensors to detect the loose wheel? Would it mindlessly continue driving with a wobbly wheel, like an idiot? Through the canyon, with tall cliffs.
Will the camera be able to see in swirling snow? My eyes get tired, their CPU may have a meltdown.
Intelligence without senses can be similar to insanity.
A.I. applied to a game of Chess is 'trivial' (relatively) because the I/O is well defined, limited and easy to implement. A.I. of the sort used with IBM's Watson machine playing word games is equally "trivial" for the same reason. But designing a robot vacuum cleaner to recognize a sticky dog turd is apparently something that is beyond human capability at this time.
Famously, "A.I. is hard.", even in limited sandboxes. In the real world, it's way, way, way more difficult than that.
These self-driving cars are going to spawn their own TV Channel. They're going to get themselves into all sorts of trouble, including dragging screaming humans, people that somehow half fell out of the car, mindlessly dragging them for miles and miles down the freeway.
They should name these new self-driving cars 'Christine', after the grusome movie.
Fetch your chair and popcorn.
All of those things are things that (some) drivers already do on a daily basis. They all sound like things that could, at least sometimes, be trivially detectable by sensors and, at least sometimes, hard to detect with human senses. My kids regularly tread mud or dog shit round the house. I'm struggling to see your point beyond automated cars will get different things wrong compared to humans. Granted, there are issues of public acceptance and confidence to be considered but I am failing to see a fundamental problem
With no steering wheel or pedals, how are you going to pull up to a gas pump at a crowded, busy gas station? That's sometimes a tricky set of maneuvers, even for a meatbag. I would imagine the AI will be very conservative in such a situation, meaning the meatbags will constantly be cutting in front while the AI is waiting for a totally clear and safe path to a pump.
Or maybe these driverless cars without steering wheel or pedals run on hopes and dreams instead of gas, so never need to fill up their tanks.
Don't know where to put this and not worth a new topic.
Searched through forums to find a suitable thread and found none so here will have to do and I think a good idea.
ver watched the dash cam footage on YT?
no steering wheel? who in their right mind would get into one of those things without that and a manual/ non electrical override button.....it's total foolishness, to anybody with at least two functioning brain cells to rub together....unfortunately the vast majority of the population are total morons :-(
Biting the hand that feeds IT © 1998–2019