That should be,"Google's self driving cars can navigate drive through restaurant drive-throughs." Although I am highly impressed by Google's self-driving cars, I can't help but feel a little bit let down that the car did not navigate around tables, make way for waiters, or squeeze through tight spaces.
Actually the video doesn't show the car navigating through the drive through space, it shows it driving up to the speakers and stopping, then it cuts to a scene from the restaurant where he's ordering, then it cuts to a scene where the car is leaving the parking lot.
As a robotics junky all my red flags went off with that bit of creative editing.
Even if the car spent 10 minutes working out the close quarters navigation it would be more 'real' than the editing which makes me suspicious that the 'side seat' driver actually did that tricky bit of navigating.
Not hating on the Goog here, just really looking forward to the era of self driving vehicles but want to be realistic about their capabilities.
There's quite a bit of footage of the self driving car making its way through congested areas with people and obstacles. It was more competent than many of the minivan drivers in my neck of the woods. :)
9:00 has a good example, recognizing the state of a traffic light, other cars and pedestrians at an intersection, but lots of cool videos starting around 3:40. The whole talk is awesome and a presentation for an actual technical audience, too.
That is a great video, my observation is a bit different.
The process by which the car figures out what to do to achieve a goal requires a system for 'working backwards' from where it wants to be, and where it is. This is called 'inverse kinematics.' The more constraints you put on the path planner, the harder it is to do the plan, in fact my experience with my own robots is that the challenge goes up super-linearly at best and exponentially at worst.
In the situations in the video linked above the vehicle can often simply wait and the path options will change until there is one it can execute. But in a drive through there is a fixed route through tight constrictions where non-organic visibility is complex at best (lots of reflections / structures) and opaque at worst.
Cars that can parallel park themselves show that the problem can be solved for a given set of constraints (I actually think parallel parking is easier in this case) but the generalized solution is at least an order of magnitude above that.
Now please don't get me wrong, I have deep and wide respect for what these guys have accomplished. I want them to be successful. And solving the case of navigating into and through a drive up window (restaurant or bank for that matter) is a solid advancement in the area of self-driving transport. And making a video to show it off is a cool thing too.
Except they didn't show it.
And that is what bugged me. There is lots of video showing the car driving through traffic, and as magicalist shows video of it driving through crowded streets, and now we get a video about 'going through the drive-thru' and it doesn't show the car navigating itself through the drive-thru lane. We are left to imagine it.
Unfortunately for Google, this is a well known technique that film makers used for a shot that is either too expensive or impractical to shoot. They set up the theme, they show the characters starting toward and action, then a quick shot of them in the middle of that action, and then a shot of them exiting the action. They leave it to our fertile imaginations to 'fill in the rest.' And it is a great story telling technique.
But if you're talking about a real self driving car, and you say it can navigate these very difficult driving situations (and anyone who does robotics will immediately go "Whoa, that is a tough challenge.") then you use the film makers trick of not actually showing anything. Well its kinda like a research paper that doesn't include any supporting data. It looks like a publicity stunt and that Google is whoring out the research for some sort of 'feel good' brand buffing. I don't think that was where they intended to go with that spot.
<sarcasm>Yea, had to be faked. The blind guy obviously grabbed the steering wheel after ordering and then drove the car around to the window.</sarcasm>
Pretty awesome accomplishment. Maybe I can get one of these cars and cut back on those tickets I keep getting in my 90 mile daily commute.
"We announced our self-driving car project in 2010 to make driving safer, more enjoyable, and more efficient."
OK so Google Labs is shut down and Google starts charging for various APIs but decides the world needs to be a better place so invests into this? I'm skeptical (even though I think this is AWESOME). What's the real story here?
Is it the extra eyeball time they get by having self-driving cars? Whereas in the past people would spend 20-60 minutes, or more, concentrating exclusively on driving each day, they could now spend that time online? I guess growth at Google is difficult when you're already the most popular search engine in the world and there's a fixed 24 hour day, the best you can do is change the % of that time people spend online, and implicitly, using your search engine. No doubt the front-end will feature Internet connectivity with a Google portal.
Now I'm confused. Where did I say it can't be monetized? I quoted a statement from Google that makes it sound like they're investing in this out of the goodness of the hearts. I'm asking if anybody knows what their actual plans are for monetizing it.
How did they tell the car to stop in front of the speaker (in the drive thru)? And how did they signal to the car that they were done ordering and to go to the pick-up window? The video conveniently skips those sequences.
"a drive on a carefully programmed route" - given careful enough programming, this appears to allow for an indefinite amount of smoke and mirrors specific to this demo.
To be fair, it would be easily worth it for a blind person to spend a few hours carefully programming important routes like this. And of course, when self-driving cars are widely adopted, Taco Bell will automatically make instructions available. (This could be as simple as a standardized sign denoting the entrance and two windows of a drive through.)
> This could be as simple as a standardized sign denoting the entrance and two windows of a drive through.
3/4 of McDonalds and other restaurants I go through already have signage, road markings, etc. for the drive through. Even some of the McDonalds I've been through have signs in the windows saying "Window 1" and "Window 2" (why this is necessary for humans of an age and intellect capable of driving a car, but apparently the ability to count windows does not have an impact on motor vehicular ability).
It really wouldn't be much difficulty for the companies to standardize these signs as most already use their own anyway.
I can't tell you how many times I've been told to "drive to the first window" only to sit there for a couple of minutes before realizing they meant the second window I encounter. It may seem stupid... But better safe than annoyed, I suppose.
I just hate the "drive up and we'll bring it to you" bit.
The other option is a 2-d barcode on the road or on a sign in a position that the car can read it. That barcode leads to a URL with a machine-readable map of the drive-thru. Naturally such a sign and URL is maintained by the restaurant and the car's systems can read and understand the content.
Awesome, if you ask me.
This was discussed in a past patent mentioned on HN.
Good idea, but probably can't be relied upon given liability issues (in addition to the whole "print wrong barcode and sabotage restaurant with crashing cars thing).
Driving seems to be the only daily activity in which a large population of people need to abide by a set of logic and rules to not die and not get hit by a big financial burden in the case of an accident or ticket. Recently I've found myself wondering if society would degrade if we lost our dependance on this sort of necessary logic in our daily lives.
I not only drive to and from work, but because I'm in construction I also spend a portion of my day driving either to the job site or between job sites, often hauling a trailer that weighs over double what my truck does.
Just this morning I had a woman swearing at me and giving me the finger, with her two small children in the vehicle, because I abided by the 4-way stop outside an elementary school and she tried to run it.
Between the cell phone users, and the middle aged women in their SUV's that can't do a shoulder check or read a single lane sign, and the guy with the chronic one-upsman syndrome who guns it to 30 over the limit just to pass you on the highway and go slightly under the limit.
IMO only about 20% of people on the road actually follow the rules because they know they're right and help keep them and their passengers safe. Then there's probably another 60% who generally follow the rules simply because they don't want to get a ticket. The last 20% just don't give a shit, do what they want, when they want and god forbid they don't get it.
I'd also say there's an additional 10-30% who are purely seasonal drivers known the other 7-9 months as "cyclists". They only come out in the winter, and are likely responsible for as many if not more accidents than inclement weather. Living in southern ontario I actually dread the roads on that first heavy snow because, no joking around, I'm actually worried I'm going to kill someone and it won't be my fault.
I don't think driving is doing society any good except at providing a way for the stupid, and the stupidly unfortunate, ways out of the genetic pool.
Are you really saying that many cyclists who switch to driving in the winter are characteristically terrible drivers? This is just really at odds with my cycling experience - A biker always needs to be on the defensive just based on the fact that the average driver is likely to not even perceive a smaller vehicle. Even though I haven't seriously biked in several years, I'm still generally aware of what most cars are going to do before they actually do it. (Then again, I guess I've also seen my share of moronic bikers - the kind that think they don't really need to think about what they're doing as long as they're doing it slowly). The first snow is indeed always a mess of bad drivers, but I blame that on people forgetting how slippery snow really is.
It's not so much the snow, it's that the cyclists disappear when it starts getting near freezing, which is the exact time you get a bunch of awful drivers on the road.
I don't mean to say all bikers are bad drivers, I'm just saying the correlation between the two events (the absence of cyclists from the roadways and the new presence of masses of bad drivers) is uncanny.
There's always the bad cyclists that just scare the crap out of you, like the 60-70 year old man who's wobbling about 2 feet from side to side, essentially rendering the bike lane useless. But in my experience as a driver, I've only seen a handful - and by handful I mean one or two - cyclists that actually obey the rules of the road. I see them run red lights and stop signs, I see them go on the side walk to cut past cars to make a right turn. etc
In my experience it is the cyclists who are trained to keep a careful eye on everyone else on the road and anticipate what others are going to do, because if they do not, they die. People who are always driving are much less aware of what their high speed air-bagged (only on the inside, of course) steel tanks can do to a human body, especially if the thing they might hit is just 70kg and not a 2 metric tonne object.
Speaking as a person who has had several accidents with cars while biking and zero accidents while driving.
"Living in southern ontario I actually dread the roads on that first heavy snow because, no joking around, I'm actually worried I'm going to kill someone and it won't be my fault."
Yeah, in Calgary people lose their shit when there's a light dusting of snow... it's ridiculous.
I think your percentages and impression of safety are off. Not sure what part of the country you live in but around here(Boston) going only the speed limit on the highways will get you about run off the road resulting in all sorts of unsafe driving. In other words going the speed limit doesn't make you safe, and going 10mph over the limit doesn't mean you don't care, the safest speed to travel is with the flow of traffic.
Additionally I happily drive well over the speed limit when visiting Maine, 2 lane wide highway, nearly straight, 5 exits in 180 miles and nearly no traffic. A speed limit of 65mph isn't about whats actually safe, its about government bureaucracy (and income from tickets).
We didn't used to drive. We used to fall onto our horses drunk and trust the well trained beast to carry us back home.
Yet somehow, we built our way to modern society from those humble beginnings.
Self-driving cars are a step backwards to a benefit we previously attained gratis from animals. For instance, in remote parts of the world where mules are still used for transportation, farmers on the mountain top load their mule with the days goods then let it navigate home (e.g. 'in take') all by itself.
Mules don't get lost, run away from danger, and feed themselves on their way to the destination. They navigate treacherous terrain and need no looking after on their way. If we had a truck with the basic intelligence of a mule, it'd be a godsend to industry.
A car with the intelligence of a palfrey---enough to not run into things, and maintain a smooth ride regardless of terrain/environment----would be an amazing luxury vehicle.
"A car with the intelligence of a palfrey---enough to not run into things"
The transition into a world with only self-driving cars brings up the issue of pedestrians and other drivers with malicious intent. Your car is going 120km/h and someone purposefully jumps in front of the car; what does the car do? Or maybe other drivers would try to snake their way through traffic like an emergency vehicle by attacking this accident avoidance system vulnerability. Do we record all these instances on camera and report them automatically to police?
The car reaction time < smaller than yours it also has a better idea of it's surrounding and I bet could pick a much safer course of action than you in a .5 second time frame.
Accidents won't disappear but I do think that if a child jumps in front an autonomous car (lets say in 5-10 years) then that child has a higher chance to survive than if a human wild be driving that car.
Exactly. An autonomous car can make the decision to swerve into the empty lane next to you much faster than you ever could, if for nothing else than increased situational awareness. The sensors are monitoring a 360 degree view constantly, we are nowhere near that.
I'm interested in hearing how their system is expected to work in adverse weather conditions. I doubt that the LIDAR system would work in rain and snow, ice and other hidden ground features might be a problem, and I wonder how well their radar works in these conditions too.
Cars are definitely great because they're easy to mass produce, and require far less maintenance than a horse. We're able to drive down cost on maintenance and production with replaceable parts---something that's a lot harder to achieve in a living organism.
That's what makes cars an ubiquity today.
Of-course I'm getting off topic....
Think about it though---a car with the intelligence of a horse but far more tame. Beautiful product.
I followed one of these on US101N the other day, near Mountain View. It was one of the better drivers on the road, although it was an easy driving environment.
I have to admit I wanted to be a dick to it (tailgate, brake check, etc) to see what it would do. I wonder if anyone has crashed into one yet.
When Google wants to put these into the market, they should run ads with video of the cars in extremely dangerous situations.
Imagine watching a video of a robot car navigating coolly and safely, with split-second timing in the midst of a snowy 12 car pileup involving an 18-wheeler. That'd sell the things as a public good.
Last I have heard Google's system is not really focused on solving perception problems, especially in inclement weather. They are waiting for the research community to catch up.
The perfect conditions they drive in allow for very precise localization against a detailed (sub-cm accuracy) map of everywhere they will drive which includs pre-marked stop signs, light, cross walks, etc.
Neither can people, but computers have the ability to see through the dark, rain and snow. They have the benefit of GPS. They have the benefit of being able to work with embedded sensors in the roadway (which to be fair is in its infancy, but you know that's where it's going).
I like people a lot, but I would not bet on a person driving better than a computer. Our reaction time is way too slow and sensory input very limited.
I thought that the title was a typo, and I was imagining a Prius navigating straight on through the glass windows of a strip mall restaurant. The reality is safer, but less hilarious.
