Last week, we wrote about driverless car testing coming to California’s public roads this September 16th. Part of the approved legislation is this language:
A manufacturer shall not permit any of its autonomous vehicles to be operated on public roads in California when the operator is not seated in the vehicle’s driver seat and either: monitoring its operations and able to take over physical control of the vehicle; or, in physical control of the vehicle.
Google just released a video of their driverless car concept, tested for the first time with people outside of Google. The video is clearly a marketing piece, a bit treacly-sweet for my taste, but worth watching anyway, if only to see how far this project has come.
One thing that struck me was the lack of any form of physical override. No steering wheel, no brake pedal. Obviously, these things won’t be necessary in a driverless car, but it seems to me, this car does not meet the requirement for an operator to be able to “take over physical control of the vehicle”. Just an observation.
Personally, I would love to test drive one of these babies. The whole idea of a self-driving car fascinates me. But I think the path to driverless cars will be bumpy and, until all the bugs are worked out, absolutely requires a steering wheel and brake.
“take over physical control of the vehicle”.
You don’t need physical controls. You have a smartphone in your pocket that can run an app to let you control it. Or you can simply issue voice commands to the car like this: “Park here on the left.”, “Follow the red car.”, “Stop!” and it would stop safely on the side of the road.
You care the ‘captain’ of the car, not the ‘pilot’.
Unless something happens to the software like it shuts down or is rebooting. There could be times, especially when it is new and not completely tested, that something could happen. In that event I would want a steering wheel and other “normal” physical controls.
There’s already more software in your car than you realize.
Software that drives a vehicle is implemented and tested in a very different way than the software of a phone. Case in point, airplanes run almost entirely on software, yet you don’t hear about them crashing everyday due to bad software.
If your phone reboots, the car will just stop by the side. Something weird happening to your phone doesn’t mean the car will start spinning in circles.
Can the captain be drunk or that’s still not a good idea.
“Hey, Google Car. Let’s race that police car.”
But this software is in addition to already tested and proven software in current vehicles. I am saying that until it is proven that the failsafe or fallback to be what we know works at least for a while.
This is why we have early adopters and late adopters.
Some people won’t care about what you’re saying and they’ll adopt such tech early without having physical control over the car.
And people like you will adopt it when they decide there’s nothing scary.
And the world will keep turning.
By the way in many cars, especially higher end ones, you have no physical control over the car either.
When you’re turning the wheel, its basically a joystick, whose input drives a motor which actually turns the car.
The more you know, right?
I know how new cars work and know that much of it is sensors and motors run by software. But all that has been thoroughly tested and evolved over time. As I am sure this will be too.
But in the meantime I don’t want an early adopter stopping suddenly in front of me or rolling into me while I am driving.
I know what we’ll do. We’ll put a Google Car sensor in your car, to keep distance from Google Cars.
No, no… Then I’ll be bombarded with ads. “Now that one of our Google self driving cars has hit you, would you like me to contact Harney and Associates for you?”
Well, good job, James, on single-handedly blocking the progress of humanity.
Self-driving cars are hard enough to make, now you’re rejecting all solutions and making it a chicken-egg problem:
No self-driving cars on the road until they’ve been sufficiently road tested.
I hope you feel bad.
No chicken and egg.. just keep them off road long enough and then once on the roads sufficient safeguards. But thanks for the laugh! Single handedly! : D
This seems to follow the strong visual language started with the glass products. This style is known as fugly 2.0 beta.
“One thing that struck me was the lack of any form of physical override. No steering wheel, no brake pedal.”
Yeah, no self-destruct button, either. I mean, what if the car becomes self-aware or something? Don’t these people watch movies?
Technically it should be a “test ride”.
This is about one-and-a-half steps beyond the blood glucose testing contact lenses. It will be years before this technology goes mainstream, if ever, and if it turns out not to be a good ad sales platform Google may strangle it in its crib like so many other products. I guess if you can’t make news with your “real” products, you’ve got to make news with your vaporware.
The application has so many possibilities, watching the video you see a blind man that is able to “drive” the car and get to a destination. You see people in their retirement years who might not be driving based on reflexes and possibly limited vision that are able to “drive” the car. Not to mention the people that text, email, eat and everything else they do while driving a car with a steering wheel. We aren’t there yet but its cool to see where its going. Imagine a person in a wheel chair that is able to roll in the car, lock the wheels and go.
odi – are you serious? your google car is about to drive into a train, and youre going to fiddly-fart around for your (locked) smartphone, authenticate, launch the app, and interact with it…..and youre telling me thats faster and better than yanking the wheel?
riiiight.
john – yet for all that software, the pilot can grab the stick any time he wants. oops.
JV is missing the point. no one is claiming that cars/planes dont have software. what people are asking for is a way to quickly give overriding instructions to the vehicle — via manual controls.
launching an app to do this instead is time-consuming and stupid.
CA law requires the passenger of a self-driving car to be able to take over in case of stupid. and yes, stupid happens.
there is absolutely nothing faster than grabbing physical controls, whether those controls use software under the hood or not. fiddily-farting for an app to given override instructions is insane.
feel free to turn that into any strawman you wish….
Ok, can you give an example of a situation, where a computer would do a worse job than a human with physical controls in front of himself.
Please refrain from referring to Godzilla approaching and other similar scenarios.
They said it will have controls in the future. These prototypes included one feature: autonomous driving.
So…it’s coming.
Targeting 2018.
Salma Hyek & I are targettting 2018 for our wedding. Let’s see which one actually happens.
You’re acting like these things are fully-baked, ready to go products. We’re talking about TEST DRIVING here. Why are they test driving? Because the entire system is NOT fully baked ready to go.
So maybe in the future when these things are ready to go and completely tested then yes, a computer will do a better job driving than a human. But as of right now when the system is still a PROTOTYPE in testing, you are a fool for having no caution whatsoever. If one of the sensors on-board malfunctions, the computer no longer has eyes. In those instances, it would absolutely be more beneficial having a steering wheel and brake pedal versus panic-grabbing your smartphone from your pocket, unlocking, finding the app, and then using an app to handle the emergency.
And how would you handle said emergency? Steer around and hit the brakes?
The computer can panic for you too.
The law says other wise. They don’t want to risk someone dying because the system couldn’t understand your words or your phone was dead.
That’s not what I said. You may never need to use your manual controls, but say you want to manually guide the car out of a ditch or something, don’t tell me that warrants a steering wheel and pedals in the way the other 99.9% of the time. I can’t believe the lack of critical thought. You won’t need a steering wheel the rest of the time because, guess what, it is a self-driving car.
Oh, and I love, having the actual implementation selected by Google on my side. Google didn’t put a steering wheel and pedals in the car. So even if the smartphone idea is stupid, at least I know you are wrong about having to need physical controls. Voice or something else would be the “manual override” if any. The car won’t run into a train, silly.
I vote for manufacturers including autonomy. That’s way more likely.
But, beyond your snark, Google said that’s their target for production to begin. We’ll see if they’re able to get someone to sign on.
Since you guys keep commenting and arguing about being able to take control, please read the comment I’m commenting on. Physical controls are coming. Y’all are arguing a moot point.
As usual, Google is blowing its wad in the wrong direction and making me wonder what problem they’re trying to solve. If they are trying to make “driving safer” this will be a failure because their cars will never replace all (or even most) of the cars on the roads of the world. What they should be doing is making a “third party add on” to other cars that will help drivers avoid accidents. The only problem these cars solve is how to feed ads to passengers while they’re transported short distances. Instead of safer roads, they’re making ad driven taxis to fill a niche market, which isn’t the problem that needs to be solved.
Realistically, I don’t think having a steering wheel and brake will help. Most people will realize it’s made a critical error when they hear the crunch, or at least when jamming the brake doesn’t stop the car in time.
Fun to see my good friends Nick and Judy in the video!
Do you really think that self-driving cars will not eventually replace most or all other cars?
How many manual transmissions do you see these days? How many horse-drawn carriages?
Less traffic, fewer accidents, eliminating drunk driving, making blind/disabled/elderly people mobile… there are so. many. clear. benefits.
I’m sure they will, EVENTUALLY, but that is a very LOOOONG way off. If Google really want’s to capture that market they should take the first logical step of making it easy for other car makers to adopt their system as the default safety and GPS controls which could begin saving lives within the next year. How many people do you really think will trust the makers of Android and Google Glass with their lives? Especially without even a steering wheel or breaks in case of an emergency? Not me (yet), and I am far from a Luddite. Planning for the magical future is great for the people who can afford it, but teaching the present to evolve toward that future is far more practical and useful for far more people.
Ugh…y’all are killing me with the “no steering wheel” junk. It is a prototype BEFORE the controls are added.
And, as usual around here, your “Google should be doing ___” statement is exactly what they’re doing: http://www.openautoalliance.net.
How is that the logical next step? Should Tesla have tried to convince other car makers to adopt their electric engine? Does anyone really think they would have been more successful that way?
Considering that Android is the most popular mobile operating system, I don’t see why people would distrust Google with their lives, and I certainly can’t see anyone caring enough about Glass for it to sway them one way or the other.
Also, re: steering wheel and breaks, there is no way a human behind the wheel will be safer than a computer.
No, that is fundamentally untrue, given the current state of technology, and the state of technology in the foreseeable future. This driverless car can’t deal with hundreds of common driving situations that are not fixed just by pressing a brake (whether with the foot or by saying “stop.”) If a child runs out in front of this car, the child gets squashed. Period. If this car hits an oil slick on the road, you are done.
This is just a car with auto pilot and the driver’s controls removed. Notice that planes have had auto pilot for many years and they all still have controls. In spite of having tens of millions of dollars worth of equipment, much more sophisticated onboard computers than this car, and flying in air traffic lanes with controllers and other help from the ground.
This is an amusement park ride that Google is using to perform a marketing stunt.
You also don’t hear about planes with no controls. And planes have had auto pilot for decades.
Cars are not only dangerous to the person who is driving them. They are dangerous to everyone around them. This video features a vehicle that does not even qualify as a car. It’s an amusement park ride.
We already have driverless cars. They’re called “buses” and “trains” and “cabs.” If you want to travel around without driving, that has been available for a long, long time.
This is just a gee-gaw for nerds. It’s way, way too unsophisticated to actually fulfill its promises.
The primary problem with cars is not replacing the driver with a computer that can only handle a small minority of driving situations, and then, only in great weather (notice these cars are tested in California and Arizona.) The primary problem with cars is the destruction of the environment and harm to people’s health that is created by the internal combustion engine. Way, way more people are killed by the internal combustion engine than are killed by driver errors, and driver errors are not even fixed by this driverless car which can’t handle many common driving situations even as well as a human. And it is not just the emissions from the engines — there is also the damage to the environment from oil extraction and transport. And there is also the wars and oppression associated with oil production. So when it comes to advancing cars into the future, that is being led by Tesla, not Google. If you’re excited about this driverless car, you have your priorities wrong.
This is just an example of Google’s lack of connection to the real world, which comes from the easy money they are making from their advertising monopoly. They don’t have to actually solve any real problems in order to get paid, same as Microsoft in the late 1990’s. So their marketing is all science fiction.
The point is: you can’t remove the driver’s controls unless/until the car is a better driver than a human driver. That is absolutely and fundamentally not the case with Google’s cars. They cannot even drive in rain or snow.
All these cars have today is a basic autopilot, the kind of thing a driver could engage after driving himself/herself onto a non-crowded highway, in good weather conditions. Like an interstate in Texas.
This is just a Google marketing/branding rah-rah video. It’s science fiction.
And this video takes advantage of one of the worst and most mistaken tenets of science fiction: that computers are infallible and humans are fallible. False. Computers are programmed by fallible humans, and typically, very fallible ones, with limited understanding of anything other than computers.
So one marketing point Google uses is they say that a self-driving car has gone 100,000 kilometers without causing an accident. Well, so have most human drivers. Most car accidents happen because of bad weather, which would also cause this driverless car to careen off the road. Humans are actually great drivers. Much better drivers than computers right now, and for the foreseeable future.
The one thing this driverless car gets right is that most of the car dashboard should be removed, in order to stop distracting the driver. But the steering wheel and accelerator and brake pedal should not. Because without them, this car cannot fulfill most of the functions that are expected from a car.
This driverless car cannot fulfill 99.9% of driving functions. You are basing your opinion on bad data and wrong facts.
You want a steering wheel and brake pedal WHEN IT RAINS. Not just when your car is in a ditch. You want a steering wheel and brake pedal about 90% of the time. The situations that the auto pilot can handle are a small minority of really, really easy driving conditions. The kinds of conditions where the driver has to fight boredom, like a long highway drive in the desert.
And how did your car get into a ditch if the auto pilot is so great you don’t need a steering wheel? Think about it for a second.
Total strawman argument.
No car today has a self-destruct button. But all cars can drive in rain, snow, or swerve to avoid a child that runs out into the street. But not the Google car. The Google car auto pilot can’t do any of those things.
THAT is why people are saying it is ridiculous to promote the idea of removing the steering wheel.
Driving in rain. The Google car cannot drive in rain.
Driving in snow. The Google car cannot drive in snow.
Swerving to avoid a human or animal who has run out into the road. The Google car cannot do that.
COMPUTERS ARE NOT INFALLIBLE.
You’re using an antique 1950’s idea of perfect computers that don’t make mistakes. Computers are not gods. They are not even as smart as humans. They are just a way to store human thoughts and play them back later. Every situation that this Google car encounters has to be “pre-driven” by a programmer at Google.
So this is not a “self-driving car” — it is a Google-driven car. A HUMAN IS STILL DRIVING IT. But instead of the human being behind the wheel of the car, the human is behind a desk at Google.
“But instead of the human being behind the wheel of the car, the human is behind a desk at Google.”
^ Now that’s just stupid.
[meant to post this the other day]
Based on the latest most valuable brand results, Google went up 40% and Apple dropped 20%: http://www.tuaw.com/2014/05/21/google-bests-apple-as-worlds-most-valuable-brand/?ncid=rss_truncated.
I think plenty would be happy with the makers of Android and Glass protecting them.
It’s a cute idea, and if Google sticks with it, I imagine they’ll address a lot of its issues. But I’d rather have improved public transportation and fewer cars on the road.