Apple 'Accelerating' Efforts to Catch Up With Self-Driving Competition, Expands California Fleet to 27 Vehicles

[belsokar]

You really think programming a plane to fly from LAX to JFK is MORE complicated than programming a car to take you from your home to the grocery store? Not even close. Not even in the same ballpark. Not even the same game!

The number of unquantifiable variables in the trip to the grocery store for your car is so far greater than flying a plane from one airport to another that the comparison is ludicrous. Have you really thought this through?

Well it's amazing how many pilots and aeronautical engineers we have on this board! The reality of flight, is that it is so much more complicated than rolling an object on a flat surface, that trying to compare them is indeed apples and buffalo...only most of you have it completely wrong. Imagine trying to program software to land an airplane in shear wind conditions. Is wind predictable? Is it something a sensor can just "see" ahead? It's constantly shifting in strength, speed, and direction...and advanced software and equipment learn to adjust and still land your plane just fine. Just because the majority of you already "trust" this software to control your safety, doesn't make it any less complicated than autonomous cars.

Trying to compare a trip to the grocery store vs flight is very funny! Again, wind variables alone present an infinite amount of TRULY "unquantifiable" variables. And it is truly infinite, because no shear winds will ever be the same.

In a plane, there are engines that need to be considered, there is a little thing called "lift" that will keep you up in the air...there are things like turbulence...you have things like fuel consumption, constantly changing aircraft weight...constantly shifting weight within your aircraft...electrical systems...pressure...all things controlled by software, that decide if you and your family will live or die...

The big difference is time, trust, and economies of scale...the systems used to keep us safe in the air will eventually keep us safe on the ground...the algorithms and software will make automobile accidents a thing of the past...or so rare that we may actually report them in the news when they happen...years, followed by decades, of research and fine tuning and shrinking more affordable technology will make autonomous vehicles so mundane, that you won't think twice about it...just like you don't think twice about it when you get on a plane...

I will say though, that those of you trying to debate that cars are more complicated gave me a good laugh today...for that I thank you.

 

[IJ Reilly]

None of that makes flying more complicated to automate.

All of the parameters in flying that are important are accurately and reliably measured and reported to the computer systems. Sure, the software has to be complex to handle the data, but the plane doesn't have to be able to read the landscape beyond what its sensors can measure.

Altitude, GPS position (which is used for ground velocity), environmental conditions, air speed, pressures, level, etc... all use very specifically purposes and well sorted sensors to report their data. The plane doesn't have to do any guess-work to figure out what's going on, because airspace is entirely open with the exception of other planes and high-rise buildings (both of which are accurately known to the plane).

Everything the car encounters is unpredictable and there is no accurate way for the car to see and interpret what's going on with accuracy. That's why they rely on a ton of cameras, they're visually looking and then trying to figure out what that data means. A plane can precisely know all its parameters at once and the programming just does the work. The car doesn't.

Once the automated car is in a snowstorm or construction zone, it loses much of its ability to read the situation around it.

Like I say, getting cars to accurately "see" everything is a colossal challenge that will need to be overcome.

They are both complicated, for unique reasons. You may not realize that legally airplanes do not have to carry more than very basic mechanical instrumentation, not even a radio or an electrical system. These airplanes share the airspace with aircraft that carry massive amounts of instrumentation and navigation equipment. They move around the sky at vastly different rates of speed and in every possible direction, including up and down. You should not underestimate the complexities of the 3D-aspect of aviation.

Pilots tend to rely on the "big sky theory" to avoid collisions but what this really means is, look, look, look, and then, look some more. No matter what ATC tells you or doesn't tell you. Terrain is also a big factor. It's pretty well mapped in terms of topography, but then, lots of obstructions close to the ground are not mapped at all, and not (such as wires) easy to detect.

Neither one is inherently easier than the other. They are different problems with their own variables.
[doublepost=1516920011][/doublepost]

You really think programming a plane to fly from LAX to JFK is MORE complicated than programming a car to take you from your home to the grocery store? Not even close. Not even in the same ballpark. Not even the same game!

The number of unquantifiable variables in the trip to the grocery store for your car is so far greater than flying a plane from one airport to another that the comparison is ludicrous. Have you really thought this through?

Yes.

 

[wbeasley]

I would believe Apple were making a car if I saw a report Samsung were building one too

 

[G5isAlive]

If I am honest, right now I don't want to be driven by a car controlled by Apple's software.

have to agree, I am almost the definition of Apple fan boy, but given their 'success' with Maps, and their foray into AI with Siri, this might be Apple technology I will pass on. Its safer to sit home watching Apple TV!

 

[LordVic]

I would believe Apple were making a car if I saw a report Samsung were building one too

Samsung has been making automobiles for decades. (I believe 1990s)

 

[KazKam]

Well it's amazing how many pilots and aeronautical engineers we have on this board! The reality of flight, is that it is so much more complicated than rolling an object on a flat surface, that trying to compare them is indeed apples and buffalo...only most of you have it completely wrong. Imagine trying to program software to land an airplane in shear wind conditions. Is wind predictable? Is it something a sensor can just "see" ahead? It's constantly shifting in strength, speed, and direction...and advanced software and equipment learn to adjust and still land your plane just fine. Just because the majority of you already "trust" this software to control your safety, doesn't make it any less complicated than autonomous cars.

Trying to compare a trip to the grocery store vs flight is very funny! Again, wind variables alone present an infinite amount of TRULY "unquantifiable" variables. And it is truly infinite, because no shear winds will ever be the same.

In a plane, there are engines that need to be considered, there is a little thing called "lift" that will keep you up in the air...there are things like turbulence...you have things like fuel consumption, constantly changing aircraft weight...constantly shifting weight within your aircraft...electrical systems...pressure...all things controlled by software, that decide if you and your family will live or die...

The big difference is time, trust, and economies of scale...the systems used to keep us safe in the air will eventually keep us safe on the ground...the algorithms and software will make automobile accidents a thing of the past...or so rare that we may actually report them in the news when they happen...years, followed by decades, of research and fine tuning and shrinking more affordable technology will make autonomous vehicles so mundane, that you won't think twice about it...just like you don't think twice about it when you get on a plane...

I will say though, that those of you trying to debate that cars are more complicated gave me a good laugh today...for that I thank you.

I think what you and the poster I originally responded to are confusing is the complexity of flight vs the complexity of driving... all things being equal. But that's the problem, all things are not close to equal in those two scenarios.

Programming a plane to fly in a straight line (from one highly controlled and monitored point to another) and a car to drive in a straight line (form an infinite number of points to an infinite number of points) only taking into account he physics involved either in a perfectly clear sky or on a perfectly flat surface, respectively, are both complex tasks. However, that's the relatively EASY part. In both cases.

Where the wrench gets thrown in is not in the complexity of the physics, but in the complexity of the ENVIRONMENT. Take your mind up into the sky a few thousand feet. Look around. How many external environmental variables are you having to account for, besides the weather, obviously, and the occasional airplane that also has three dimensions of space to be located in at any given time? Now go stand in the middle of the street. See the difference?

 

[TrentS]

The "human ego" and the "human temper" and "human error" are what usually causes most accidents. AI will most likely not have any of these in the mix ( unless some bozo programmer forgets to place a decimal in the wrong place or something ).

Imagine if they program "ego" and "rage" into the AI!!

 

[wbeasley]

I'm not in the market for a new car. I actually love my 2012 SE. the styling fits what I like (Clean lines with a wedge shape, not curvey, very nascar like) and is sufficiently powerful enough for a comfortable drive (obviously it's not a spots car at 180hp). PLus I have only 40,000km on the car as I only need to drive casually.

just using the anecdote that they can't even get me to buy their latest design by giving it away at discount because of how much i hate it

Bit like the iPhone X

I could update, I'm due to update, I can afford to update but what they've delivered (late) is not floating my boat. That notch, the smaller effective screen size than my 6S+ mean I've voted with my wallet this year. If there isn't a bigger, nicer version later this year I either change the battery or upgrade to an 8+ or 8S+. Or downgrade to an SE (revamp) to keep the ease I like of Apple phones and buy an Android something for mobile media consumption. Android is a mess from fractured OS and skins. But adding SD card with bigger, nicer screens is tempting at the right price.

If you're wondering why I want/need a second device for media... I like to read in bed or listen to music/video. Even an iPad mini is too large to hold when tired. Rather not deplete phone battery either. I know, throw it on a charger overnight. But after taking off my glasses and being tired, the last thing i need is fumble around in the dark to plug in an iPhone.

Yep, first world problem...

 

[xxray]

But... why?

 

[IJ Reilly]

I think what you and the poster I originally responded to are confusing is the complexity of flight vs the complexity of driving... all things being equal. But that's the problem, all things are not close to equal in those two scenarios.

Programming a plane to fly in a straight line (from one highly controlled and monitored point to another) and a car to drive in a straight line (form an infinite number of points to an infinite number of points) only taking into account he physics involved either in a perfectly clear sky or on a perfectly flat surface, respectively, are both complex tasks. However, that's the relatively EASY part. In both cases.

Where the wrench gets thrown in is not in the complexity of the physics, but in the complexity of the ENVIRONMENT. Take your mind up into the sky a few thousand feet. Look around. How many external environmental variables are you having to account for, besides the weather, obviously, and the occasional airplane that also has three dimensions of space to be located in at any given time? Now go stand in the middle of the street. See the difference?

Your "besides" factors are among the most critical complications. You cannot simply dismiss them because you don't understand them fully enough to discuss them.

 

[maflynn]

I would believe Apple were making a car if I saw a report Samsung were building one too

Ahem
Samsung makes a $300 million push into self-driving cars

 

[G5isAlive]

They are both complicated, for unique reasons. You may not realize that legally airplanes do not have to carry more than very basic mechanical instrumentation, not even a radio or an electrical system. These airplanes share the airspace with aircraft that carry massive amounts of instrumentation and navigation equipment. They move around the sky at vastly different rates of speed and in every possible direction, including up and down. You should not underestimate the complexities of the 3D-aspect of aviation.

Pilots tend to rely on the "big sky theory" to avoid collisions but what this really means is, look, look, look, and then, look some more. No matter what ATC tells you or doesn't tell you. Terrain is also a big factor. It's pretty well mapped in terms of topography, but then, lots of obstructions close to the ground are not mapped at all, and not (such as wires) easy to detect.

Neither one is inherently easier than the other. They are different problems with their own variables.
[doublepost=1516920011][/doublepost]

Yes.

you position yourself as the logical thinker.. one of these problems was solved 3 decades ago.. one of these has not been solved yet with only a guess on when it will be. Still want to take the position neither one is inherently easier?

Quibble all you want, but a self-driving car is an order of magnitude more difficult than automating flying. Its common sense. Find a pilot to disagree.

 

[LordVic]

Bit like the iPhone X

I could update, I'm due to update, I can afford to update but what they've delivered (late) is not floating my boat. That notch, the smaller effective screen size than my 6S+ mean I've voted with my wallet this year. If there isn't a bigger, nicer version later this year I either change the battery or upgrade to an 8+ or 8S+. Or downgrade to an SE (revamp) to keep the ease I like of Apple phones and buy an Android something for mobile media consumption. Android is a mess from fractured OS and skins. But adding SD card with bigger, nicer screens is tempting at the right price.

If you're wondering why I want/need a second device for media... I like to read in bed or listen to music/video. Even an iPad mini is too large to hold when tired. Rather not deplete phone battery either. I know, throw it on a charger overnight. But after taking off my glasses and being tired, the last thing i need is fumble around in the dark to plug in an iPhone.

Yep, first world problem...

Honestly for reading, I wholly recommend getting off the iPad/Iphone/phoenwhateveryou use for reading (if it's novels / simple text) and get an e-reader.

best thing I did for my head/eyes / sleep for my reading
'

 

[IJ Reilly]

you position yourself as the logical thinker.. one of these problems was solved 3 decades ago.. one of these has not been solved yet with only a guess on when it will be. Still want to take the position neither one is inherently easier?

Quibble all you want, but a self-driving car is an order of magnitude more difficult than automating flying. Its common sense.

It isn't common sense, it is just plain wrong, and I have said why in detail. I believe you aren't listening.

 

[LordVic]

Ahem
Samsung makes a $300 million push into self-driving cars

Samsung is also more likely to have more success with it since they already have automobiles and their own manufacturing to use it in.

(not speaking to quality etc)

it's why this Apple car project, whatever it turns into is just weird. What the hell are they doing. Come on apple. lift the magic sheet. Let us peek! I'm sure the investors would be curious what this money is going to
[doublepost=1516922459][/doublepost]

Bit like the iPhone X

People forget how much subjectivity plays a part to peoples emotional attachment.

"but, the notch dissapears in a while, yo wont notice it while using it!"

IRRELEVANT! i still think it's ugly! yes, I understand why it's a notch. YES< i know the practicality behind it.

I still think it's ugly! Just like I think it's ugly on essential. I would rather a uniform display that is 100% there! IMHO, a device with a 2-3mm chin and forehead that is perfectly symetrical is far more appealing than an unsymetrical design that takes out portion of the display.

 

[entropys]

I struggle to trust humans for this task. Self driving software has some way to go but eventually it will be far safer than any human driver. At that point driving a car should be forbidden, as it would be too risky compared with driverless cars.

that is so. Of course, it also means the State will have total control who, what and where on the roads.
The huge benefit of safety will conflict with the huge negative of Big Brother

 

[Altis]

They are both complicated, for unique reasons. You may not realize that legally airplanes do not have to carry more than very basic mechanical instrumentation, not even a radio or an electrical system. These airplanes share the airspace with aircraft that carry massive amounts of instrumentation and navigation equipment. They move around the sky at vastly different rates of speed and in every possible direction, including up and down. You should not underestimate the complexities of the 3D-aspect of aviation.

Pilots tend to rely on the "big sky theory" to avoid collisions but what this really means is, look, look, look, and then, look some more. No matter what ATC tells you or doesn't tell you. Terrain is also a big factor. It's pretty well mapped in terms of topography, but then, lots of obstructions close to the ground are not mapped at all, and not (such as wires) easy to detect.

Planes can read topography (to the extent that it's relevant to the plane) easily and make reliable decisions based upon it. Similarly planes can detect other aircraft. Airspace is an extremely thoroughly controlled environment compared with roads.

All of the instruments on the plane provide to the computers very reliable information. The sensors report very specific parameters about the world that the plane can use.

Cars only have limited forms of that, but mostly they generally rely on cameras. Cameras return meaningless information by themselves -- pixels mean nothing. Thus, the computers have to try to analyze them and figure out what is going on, out of nearly infinite possibilities.

Since there are endless possibilities of what the car might encounter, and what it might have to do, it's very difficult to even just extrapolate data from the images in order to determine what to do -- assuming it can still see properly at all. Cars can lose GPS signal (underneath structures) and cameras can be blocked. Roads can have detours along unpaved sections. Might have to drive around an unusual obstruction. Cars, pedestrians, cyclists, and other highly unpredictable obstacles can present themselves out of nowhere, and some can be very difficult to see and detect. You get the idea.

Neither one is inherently easier than the other. They are different problems with their own variables.

Incorrect. Planes themselves are much more complicated than cars, but the ability for technology to accurately read the parameters of the world around the plane are more straight-forward.

People tend to think that because flying itself is more complicated, it must be more complicated to automate. But it's the environment that generally dictates the complexity, the variables -- getting the data. Once you have all the data, the rest is programming.

 

[mvray]

The updated flux capacitor looks sweet.

 

[IJ Reilly]

Planes can read topography (to the extent that it's relevant to the plane) easily and make reliable decisions based upon it. Similarly planes can detect other aircraft. Airspace is an extremely thoroughly controlled environment compared with roads.

All of the instruments on the plane provide to the computers very reliable information. The sensors report very specific parameters about the world that the plane can use.

Cars only have limited forms of that, but mostly they generally rely on cameras. Cameras return meaningless information by themselves -- pixels mean nothing. Thus, the computers have to try to analyze them and figure out what is going on, out of nearly infinite possibilities.

Since there are endless possibilities of what the car might encounter, and what it might have to do, it's very difficult to even just extrapolate data from the images in order to determine what to do -- assuming it can still see properly at all. Cars can lose GPS signal (underneath structures) and cameras can be blocked. Roads can have detours along unpaved sections. Might have to drive around an unusual obstruction. Cars, pedestrians, cyclists, and other highly unpredictable obstacles can present themselves out of nowhere, and some can be very difficult to see and detect. You get the idea.

Incorrect. Planes themselves are much more complicated than cars, but the ability for technology to accurately read the parameters of the world around the plane are more straight-forward.

People tend to think that because flying itself is more complicated, it must be more complicated to automate. But it's the environment that generally dictates the complexity, the variables -- getting the data. Once you have all the data, the rest is programming.

Pretty much everything you said about aviation is completely inaccurate.

Airplanes can "read" none of these factors. They cannot "detect" other aircraft. The majority of the national airspace is in fact not controlled, leading to the other fact, which is that most flight is not controlled. Most airplanes in fact have little to no ability to sense anything and many fly with virtually no information or computational ability whatsoever beyond what the pilot can see, and fairly primitive instruments that are far from precise. Flight takes place in a three-dimensional, highly variable and often very crowded environment, with little room for error, and none for failure. These are simply facts, whether you care to know about them or not. The people who tend to think flying is complicated are called pilots. But what would they know?

Note that I am not saying the two-dimensional environment for cars is uncomplicated. I am saying (once again) that these two environment are complicated in different ways.

 

[falainber]

So is this just going to become a part of Car Play? I can't wait to see how much BMW charges yearly for this!

Knowing Apple, it is going to become the part if a phone (3D touch + 3D camera + autodriving)

 

[Soccertess]

This could be Apple's next game changer. I'm bored of slightly thinner iPhones, time for something innovative.

Apple's next game changer? Apple hasn't released anything game changer for a long time (RIP Steve)

I do like the Apple watch, just not revolutionary as Tim Cook predicted. The w1 chip was interesting, but so limited that hardly a world game changer.

If anything, just like the article said Apple is playing catch up up to the rest of the world. I can't think of any sector where Apple is leading (except on price). Apple holds huge market share in USA, but rest of the world is dominated by android.

 

[chabig]

Huh... while I agree with your sentiment, for me, it goes beyond that.... I don't want to pay for Apple's car driving software.

Don't worry. Google's cars will drive you for "free" while they track your location and subject you to non-stop advertising.

 

[Altis]

Note that I am not saying the two-dimensional environment for cars is uncomplicated. I am saying (once again) that these two environment are complicated in different ways.

Cars also operate in a 3-dimensional environment. They have to take grade into account, can go up/down layers of roads (such as parking garage) that give the same map location, have to observe elements high and low, etc.

Pretty much everything you said about aviation is completely inaccurate.

Airplanes can "read" none of these factors. They cannot "detect" other aircraft. The majority of the national airspace is in fact not controlled, leading to the other fact, which is that most flight is not controlled. Most airplanes in fact have little to no ability to sense anything and many fly with virtually no information or computational ability whatsoever beyond what the pilot can see, and fairly primitive instruments that are far from precise. Flight takes place in a three-dimensional, highly variable and often very crowded environment, with little room for error, and none for failure.

In terms of complexity, ground space >>>> air space; the latter is exceptionally more predictable.

That's why planes have been using autopilot as the norm for over half a century, well before modern computers were available, while even present day cars have the most limited automation. It's also why planes can fly themselves but won't taxi on their own.

Just look at the colossal array of cameras and sensors on the car in the OP. It's nuts. And that has to be combined with the most advanced algorithms to try to interpret what's going on on the road around it.

Of course the guidance system on planes can "read" the flight parameters. That's what it uses to pilot the plane automatically. I have no idea how you can pretend they don't -- that's how automation works. Planes can absolutely detect one another with systems like TCAS. For other collision precautions, look to how heavily controlled airspace is, and the presence of pilots still.

These are simply facts, whether you care to know about them or not. The people who tend to think flying is complicated are called pilots. But what would they know?

Nobody is saying flying isn't complicated, of course it is. That's why they still have pilots on planes. But the physical control of an airplane is fairly straight-forward to automate since there isn't much unexpected to account for. Getting the plane in the right place, velocity, altitude, etc, are all perfectly manageable.

On the other hand, interpreting camera images and trying to figure out from the countless number of scenarios what in the world is happening... that's an exceptionally complex challenge. I say this as someone who works in commercial automation.

Your nasty attitude towards others makes this seem a worthless endeavor.

 

[AFDoc]

Am I the only one that doesn't want to be sharing the road with self driving cars?

 

[bernuli]

HomePod doesn't even have stereo channel sound ability and they are working on a car?...that drives itself?