What a difference a couple of weeks can make. We knew on February 5th that Apple was offering quarter-million dollar signing bonuses to Tesla engineers to persuade them to jump ship, but the idea that the company planned to make a car was just a vague rumor. Fast-forward a fortnight and it’s now being treated as established fact.
Our own exclusive reporting on the sheer range of automotive hires by Apple makes it clear that the company is, at the very least, seriously investigating the possibility, with a 1,000-strong team reportedly approved by Tim Cook. And while we need to bear in mind the cautionary note in Seth Weintraub’s piece that there’s a big difference between an RD project and a real, live product, at this stage an Apple car seems more likely than not.
But if Bloomberg is right that Apple plans to launch a car by 2020, I think it’s important to recognize what form that car will and won’t take (spoiler: it won’t look like the above) …Â
Let’s start with the obvious things we can expect from an Apple Car.
Of course, it will be electric. That’s a no-brainer. It’s the direction in which the industry is headed, is fully in line with Apple’s environmental commitments and Tesla has more than demonstrated that electric cars can be every bit as exciting as their dinosaur-burning competition.
Of course, it will be beautiful. Whether Apple is targeting a sleek sportscar or practical people-carrier, the styling will be gorgeous.
Of course, it will be hi-tech. Under the hood, it will use hi-tech battery technology. Apple will, as Jeremy Horwitz argued, solve the mess of the center stack, giving us a beautiful control panel for all the in-car amenities. Perhaps it will bring its user-interface expertise to bear on some of the still-clunky controls beyond that center stack, too.
And of course, there’ll be an app for that. Tesla has already demonstrated just how beautifully you can integrate an iPhone app with the car to do everything from monitoring battery-charging to ensuring the cabin is at the perfect temperature before you get into it.
It’s also a no-brainer that Apple is working on a self-driving car. That’s the most exciting upcoming development in the automotive world, and there is no more obvious place for Apple to find the intersection of technology and the automotive arts.
Fully autonomous cars will be a fantastic thing. You’ll be able to get into your car after a long, tiring day, press the auto-pilot button, have your car ask you where you want to go, say “Home†and then spend the journey sleeping or browsing 9to5Mac on your iPad Holo (I should probably trademark that now).
But that brings us back to this 2020 business. While the progress made in self-driving cars in recent years has been incredibly impressive, there remains a vast gap between what we have seen to date and the scenario I’ve just outlined.
The answer to that lays in how Google’s autonomous cars work. First, the vast majority of those miles were driven on freeways. This is the least-demanding driving task possible. There are clear lane-markings, few junctions, no pedestrians or cyclists, no railroad crossings … in short, freeways are Roads Lite.
Second, when the time came for Google to tackle the far, far trickier business of city driving, its engineers didn’t just program a computer to drive a car and then let it loose on the streets. They instead started out by creating, literally, an inch-level 3D map of a specific series of streets. Every pothole, every road sign, every painted line on the road surfaceâ€"all of those details were fed into the computer model used by the car. When anything changes, the street has to be remapped before the car can drive along it.
With all of this detail in place, the car is an incredibly capable driver. It knows, when the model tells it that it’s reached a railroad crossing, that it must wait until the far side of the crossing is clear before it proceeds. It knows, when turning right, that it should allow pedestrians to cross first. It knows, when it sees a cyclist raise their left arm, that it must wait behind and allow the cyclist to change lanes.
But all of this is possible thanks to that constantly-updated, inch-level computer model of every street on which the car drives. Even in Google’s home town of Mountain View, Google has only mapped a fraction of the streets. Rolling that out to every street in the whole of the USAâ€"all 3,980,817 miles of themâ€"would be a massive undertaking. Keeping every single one of those miles updated every time anything changes would be an even bigger one.
Of course, eventually autonomous cars will be smart enough not to need this mapping. They will be able to use general rules and pattern-recognition to figure things out for themselves. But that level of technological sophistication isn’t here today, and isn’t going to be here in five years time. Probably not at all. Definitely not working, tested and sufficiently reliable to release into the wild. And absolutely not approved for public use by the regulatory authorities.
Now, that doesn’t mean the Apple car won’t proudly proclaim itself to have self-driving technology. There are cars today that make that claim, like Tesla’s “Autopilot.†But what that term means today is very far removed from that delicious vision of kicking back and letting the car drive us home. So, what sort of capabilities exist today?
Self-parking cars. Pull up next to an empty slot in a parking lot, or a parking space on the street, and the car can drive the last few feet into the space. Tesla is taking this a stage further, allowing the car to learn its usual parking spot at your home or work. You can get out of the car and leave it to park itself.
Lane-following and lane-changing on freeways. For years, we’ve had technology that allows a car to recognize clear lane markings and remain within that lane, adjusting speed to remain a safe distance behind the vehicle in front. We’ve more recently seen the logical extension of this: changing lanes to overtake a slower vehicle, moving back into the original lane afterwards.
Emergency braking. Carsâ€"and even trucksâ€"can spot an obstacle in the road and apply the brakes, bringing the vehicle to a halt to avoid a collision. Volvo has taken this a step further and programmed cars to recognize pedestrians and cyclists heading toward a road, not waiting until they actually move into the car’s path.
Speed control. Some cars can recognize speed limit signs and automatically adjust the speed of the car to comply.
Tesla demonstrated a number of these things last year. The executive summary is that today’s ‘self-driving’ cars can do a small number of very specific things well enough to form extremely useful driver aids, but they don’t come anywhere close to replacing a human driver.
So, what can we expect from Apple? Well, technology will move on, of course. Apple and others will continue to build on the existing capabilities described above, so an Apple Car will be able to do more than cars can do today. But I don’t think it will push the technology much further than other manufacturers, and I can explain why in one word: iPhone.
The iPhone revolutionized the smartphone, but it didn’t do this by adding much in the way of new technology. Touchscreen smartphones had existed for years before the iPhone. Multi-touch was really the only significant technological innovation. What made the iPhone such a ground-breaking product was not that it did anything new, but instead that it did it so well.
Smartphones were, at the time, for geeks. Apple turned them into a mass-market product. Existing manufacturers had clunky-looking smartphones; Apple created a beautiful one. Other smartphones had to be controlled with a fiddly stylus; Apple enabled us to use our fingers. Others had physical keyboards; Apple created a usable on-screen one. Others had UIs designed by geeks for geeks; Apple created one so simple that non-techy consumers could use one without instruction.
That was Apple’s strength then, and that is Apple’s strength now. Watch technology emerge. Watch what other companies do. Wait until the time is right, then do the same thing, only better. It’s what the company has done with the smartphone, the tablet, the smartwatch … and is what Apple will do with the smartcar.
So, if that 2020 date is right, that’s what I think we should expect from an Apple Car. Electric. Beautiful. Hi-tech. Great user-interface. A companion app. Some cool driver aids, probably using that self-driving terminology. But not, in five years time, anything close to a fully automomous car.
Main image:Â Rinspeed (XchangeE concept car based on Tesla Model S). Google car: NBCNews. Lane-change: Tesla. Final image: Volvo.
Bloomberg: Apple planning to launch its own car by the year 2020
Poll: Do you think Apple should actually build an electric car?
An interesting technology which is (slightly) related but (very) cool is that used by London buses. Their exact location in relation to each other and traffic signals is known, so that priority can be given to them at junctions, shortening red signals, for example.
If you had two or more Google self-driving cars in close range, their LIDAR sensors would cast lasers into each other, polluting the readings beyond usability. The only foreseeable solution is a TDMA scheme for firing your laser. This definitely doesn’t scale to every car on the road using these sensors.
Without this magic sensor, we’re left with basic cameras as our fanciest device to use, and there just hasn’t been as much progress with that technology to get us anywhere as close to full autonomy as Google has with LIDAR.
The cars were checked but no error or defective part was found.
According to Swiss technicians the blocked electronic occurred to electro smog from all the antennas in the tunnel (mobile phones, radar cams, surveillance cams, radio antennas etc.) all the different frequencies from these antennas are reflected like a parabol antenna to the cars electronic which just couldn’t deal with all the data. Things are even worse, by multiplexing a lot of data over a single cable, car makers avoid having 1000 of km cables in cars for every sensor. Accessing and fetching the right data in milliseconds is a horse-work for any computer. If the multiplexed data is being infiltrated by foreign signals (as in the Gubrist Tunnel) any electronic runs Amok.
(articles in German language:
http://www.autobild.de/artikel/elektrosmog-37425.html,
http://www.emf.ethz.ch/archive/var/sb_gysel_pref5.pdf (page 46))
Sorry to disagree, but I guess that if Apple does not come with a fully self-driven car by 2020, it will be way behind other car makers.
More info here: http://www.autoblog.com/2015/02/19/volvo-self-driving-cars-2017-official-video/
At least the video suggests the car depends more on the sensors, cameras, lasers and etc. than on the mapping system.
They might roll out an electric car with GPS and all the bells and whistles, and have that car have a camera on each side of it,that will start mapping everyting around it, as soon as you go faster than 5 miles per hour, so wil even 10,000 cars sold, they will almost instantly have a good road mapping from all pver the country. Google does that with GPS and your phone. search for “location History†on google, which will show your precise location every 15 min of the day. Whats to stop apple to do this with cars and cameras in them, so that they can collect as much data as possible, and by release date of Apple Car 2.0 make it self driving, as it will have all that massive mapping, inch by inch, consistently updated on road conditions, army of apple cars driving the streets.
You only state Google and Tesla. But VW, Toyota, Audi, Renault etc. all have self driving cars since years, but they don’t make the hype of being on road anytime soon.
Why? Because the legislation is way too complicated to change to self driving computer decisions. The last decision should be done by people not a computer. I know the USA is computer crazy and they believe every decision made by computers (why do they trust computer enabled elections?, why do they trust opinions on Facebook or other social networks? Why do they trust computer decisions to catch (non-)suspects (FBI, NSA) or kill people (CIA drones)…)
A simple example: the driving laws in Europe say, that if an animal (dog, cat, deer…) crosses the street and you don’t have the possibility to break or dodge the animal, you have to run over the animal (even if that means killing it or damaging your car.)
Will Google car avoid the animal and hit the next tree? How can it make a difference between a deer and a human being with a stroller? During every bigger storm hitting Europe, people are killed by falling trees. How will a self driving car react? Full-Brake or accelerate to drive away from the falling tree? (In Terminator this may work, but our computers can’t make these decisions yet)
Plus millions of km of roads are missing signs, have sun bleached signs, are missing lane markings etc. and a lot of people don’t behave according to the street rules (don’t stop at red lights or STOP signs…) how does the computer software act here which only works for right or wrong decisions (binary system, not perhaps or in between) .
In theory everything works great and tests do prove it on prepared roads or roads with limited interactions. The technic works already but only in a perfect world.
You only have to take the new visual systems in modern cars: self parking or recognizing speed limit signs …
It is functional but not more! The next car I want to buy will have all these electronic gadgets. But as my dealer tells me: “the sign recognition only works under good conditions, bad weather spoils it. And self parking is for those who cannot park a car. If can park, you get in any parking space the car fits in. The self parking system needs much bigger park space.†(Here it’s Europe cars, parking lots and streets are much smaller than in the USA ;-))
I really would like to see an iCar, but I think it won’t be so soon if it happens at all.
The fact that Goole is heavily investing in this technology means absolutely nothing. Google invests in a bunch of projects with most of them eventually being discontinued. The way Google operates is they require that every engineer spends 20% of their time on a personal project approved by the manager. Some of these personal projects go nowhere, but others get traction with Google investing money in the idea and building a team around it. This has been happening for a decade now. Not much has come out of this; name one Google product outside of Search/Gmail that is widely used. Name one business model outside of advertising that earns Google any significant income. There is nothing to name.
Google is a large advertising agency built by geeks who came up with a brilliant search algorithms. Google has been milking this golden cow for over a decade now by selling ads. Because geeks are still running the company, they are very much interested in investing billions in high-tech research. That’s all the self-serving car is about â€" investing billions in a research project.
As for the latest rumors of Apple building a car team, this is more likely to be the beginning of the closer integration with and future acquisition of Tesla than it is about building a self-driving car. Fortunately, Apple says “no†1,000 times for each ‘yes,’ whereas Google says “yes†almost as often as it says “no’.
On the whole no one is driving tech news as quickly and as accurately as 9to5mac. Kudos to you, and keep going strong.