Toyota Research Institute head says full autonomous driving is “not even close” • Hacker News Board

Re: Toyota Research Institute head says full autonomous driving is “not even close”

#241

Post by simonh » Mon, Aug 28, 2017, 6:11 PM UTC

AndrewKemendo wrote: post #217

kajecounterhack wrote: post #203

Earlier quoted context omitted.

Toyota and other car companies are fantastic at building cars, yes. But autonomous driving is at its core a challenging software problem. No major car manufacturer outside of perhaps Tesla has a track record of shipping software at scale. They're hardware companies. Your sentiment here is like saying Samsung is great at making phones so they must be great at Mobile OS's. They're fundamentally separate things, and we…

Toyota and other car companies are fantastic at building cars, yes. No Toyota is fantastic at designing, building, marketing, managing logistics, integrating credit systems etc...for cars. Selling a Level 5 automated car is as much marketing as it is software. It could work perfectly but if nobody trusts it enough to buy, at a low enough price point, then it won't matter.

It's hard to sell something you don't have, and Toyota doesn't have competitive self driving tech or electric car tech.

Arguing that they will be a leader in self driving cars because they are leader is selling legacy cars very much smacks of the "PC guys aren't going to just walk in" comment regarding smartphones.

There's an old saying that goes something like:

X + computer = computer

So

Teletype + computer = computer

VDU + computer = computer

Phone + computer = computer

Computers and IT driven companies are eating whole industries.

Re: Toyota Research Institute head says full autonomous driving is “not even close”

#242

Post by lb1lf » Mon, Aug 28, 2017, 6:14 PM UTC

jpindar wrote: post #204

ageofwant wrote: post #179

Earlier quoted context omitted.

Intercontinental planes are mostly autonomous. Trains are autonomous. Massive mine dump trucks are fully autonomous. These things are in use today and has been for almost a decade. They need not be perfect to be useful, they only need to be better than the average human driver. Toyota is way behind in the game. What do you expect them to say about the competition?

Not all trains are autonomous, and they have been known to crash due to operator error.

Besides, trains run on tracks, each of which accomodates one train at any given time and rarely intersect; additionally, there's all sorts of signalling and safety infrastructure embedded with the track...

In stark contrast with lane-swerving cars, intersections everywhere, temporary rerouting because of roadworks &c.

Running trains must be orders of magnitude simpler than fully autonomous vehicles.

As for planes, they have a massive benefit in being engineered and maintained for levels of reliability no car can ever hope to achieve, having an awful lot of empty air around them (not to mention being able to move in the Z plane, too, to avoid collisions.

I don't doubt there are many lessons to be learned by designers of autonomous vehicles from work already put down in the fields of trains and planes - however, I'd argue they are very different problems.

Re: Toyota Research Institute head says full autonomous driving is “not even close”

#243

Post by Pigo » Mon, Aug 28, 2017, 6:15 PM UTC

dperfect wrote: post #57

I've been saying this for a while. The problem is that people are mostly familiar with the rate of overcoming traditional technical challenges (e.g., making your current computer/phone more compact, improving graphics quality, making cameras and screens with more pixels, etc.). What people don't realize is that full autonomous driving will require more than just faster/smaller/cheaper technical innovation - it will r…

I agree with you, We all hope we'll have amazing self-driving cars real soon, it's just funny watching the responses here. If you try to say autonomous cars may not be a perfect solution and available in 6 months, or SpaceX may not have a thriving colony on Mars in 6 years, then you will ruffle some feathers. I think guys that have lived through disappointing hype before have some perspective to share.

Re: Toyota Research Institute head says full autonomous driving is “not even close”

#244

Post by projectileboy » Mon, Aug 28, 2017, 6:17 PM UTC

Something that I think is sometimes missed is that the insurance companies along with major freight haulers may drive this more than the public at large. Self driving systems don't have to be perfect; they just have to be better than humans, and that may be attainable sooner than the naysayers imagine. We're still north of 30k auto deaths per year in the U.S. alone - I'd bet money that if you waved a magic wand and replaced every car on the road today with an autopilot Tesla, traffic deaths would fall off a cliff.

Re: Toyota Research Institute head says full autonomous driving is “not even close”

#245

Post by ghaff » Mon, Aug 28, 2017, 6:27 PM UTC

throwawayjava wrote: post #223

btian wrote: post #198

Earlier quoted context omitted.

Airlines will kill to get another 10% in margin.

So are flight attendants impossible to automate/do without? Or are there perhaps other forces -- beyond technical ones -- keeping humans in the plane?

AFAIK, the FAA requires a certain number of flight attendants for safety reasons.

Re: Toyota Research Institute head says full autonomous driving is “not even close”

#246

Post by nradov » Mon, Aug 28, 2017, 6:30 PM UTC

throwawayjava wrote: post #150

nradov wrote: post #134

Earlier quoted context omitted.

Military UAVs have a high crash rate. There's no autopilot which could have successfully landed an Airbus A320 in the Hudson River.

And no autopilot would intentionally ram a plane into the side of a mountain. These edge-cases are uninteresting IMO.

On the contrary. The edge cases are really all that count.

Re: Toyota Research Institute head says full autonomous driving is “not even close”

#247

Post by devmunchies » Mon, Aug 28, 2017, 6:37 PM UTC

nayuki wrote: post #102

devmunchies wrote: post #93

Earlier quoted context omitted.

> a car misread a stop sign with a small amount of marking as a 45MPH speed limit sign That's what happens when the sign were made for humans. They should start putting QR codes under traffic signs, with the human sign as a fall back or as a second point of reference.

How can you verify at a glance that the QR Code associated with a human-readable sign is correct?

I said this in my comment: "...with the human sign as a fall back or as a second point of reference."

You wouldn't remove human sign identification from the algorithm, you just supplement it with the small QR code under the sign.

Re: Toyota Research Institute head says full autonomous driving is “not even close”

#248

Post by DoubleCribble » Mon, Aug 28, 2017, 6:41 PM UTC

speby wrote: post #236

kajecounterhack wrote: post #209

Earlier quoted context omitted.

I don't know by what metrics you're making these comparisons. Audio recognition and self-driving vehicles are apples and oranges. Autonomous driving is difficult primarily because of a changing world and low tolerance for mistakes. The state space of situations you're trying to map is just large. It will take time. But it's not even close to impossible. Audio recognition is difficult for a different reason: language…

Audio recognition may be apples and oranges to self-driving vehicles. But as it relates to autonomous driving, audio recognition (and I don't mean speech but general sounds) is something nobody has been talking about (at least publicly) as a sensor input for self-driving cars. Think about how much you rely on your hearing to drive safely and well. Sure, we use sight/depth and so forth substantially but we also rely o…

Curious, isn't it? Humans rely on auditory information to enhance situational awareness so you'd think that would factor into the design of a machine that needs as much situational awareness as possible.

For example, emergency response vehicles announce their arrival and direction often long before any visual contact is made. How will deaf autonomous cars recognize incoming emergency responders if the view is obstructed? Will they pull into the intersection just as a Fire Engine is running the red light?

Auditory input was deemed important enough that California even made it illegal to "wear a headset covering, earplugs in, or earphones covering, resting on, or inserted in, both ears" while driving just last year. [0] I suspect other jurisdictions have/will follow(ed) suit.

[0]http://leginfo.legislature.ca.gov/faces/codes_displaySection....

Re: Toyota Research Institute head says full autonomous driving is “not even close”

#249

Post by saosebastiao » Mon, Aug 28, 2017, 6:44 PM UTC

dsfyu404ed wrote: post #211

saosebastiao wrote: post #105

Earlier quoted context omitted.

> What I would like to see next is a system that alerts the driver when the light in front of them turns green. I feel like I lose 5-10 minutes a week sitting behind people who are checking their phones unaware the light has changed. I wish I could upvote this idea straight into the car manufacturers inboxes. Nothing more frustrating than having to sit through multiple light cycles cause the dipshit in front of you i…

You have one. It's called eyeballs and a horn.

It works really well when it's the person right in front of you, because you can see that they aren't paying attention. Put a few cars between you and them, and all of a sudden you don't really know the reason they aren't moving and there is a significant delay before you can infer that they aren't paying attention. It would be so much better for their own car to recognize that they are in la la land and remind them asap.

Re: Toyota Research Institute head says full autonomous driving is “not even close”

#250

Post by throwawayjava » Mon, Aug 28, 2017, 6:56 PM UTC

nradov wrote: post #246

throwawayjava wrote: post #150

Earlier quoted context omitted.

And no autopilot would intentionally ram a plane into the side of a mountain. These edge-cases are uninteresting IMO.

On the contrary. The edge cases are really all that count.

We shouldn't demand that our systems work in 100% of cases. Nine Nines -- or even less, in the case of autos -- is enough.

You can argue that autonomous flight doesn't yet have Nine Nines. I would probably agree. But "Miracle on the Hudson" was just that, and extremely rare situations should not be a standard-bearer for allowing autonomous flight. In the same way that suicidal/homicidal humans shouldn't be the standard-bearer for allowing human flight.