Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The latest chapter for the self-driving car: mastering city street driving

googleblog.blogspot.com

Post reply on HN
Print view View original

361–370 of 455 posts

Re: The latest chapter for the self-driving car: mastering city street driving

#361

Post by jfoster »

revelation wrote: post #288
dabeeeenster wrote: post #15

As a city cyclist, the Google solution looks way safer than the current status quo...

The fellow people over at Reddit have alerted me to the fact that the Google car is actually still breaking the law (while being a tremendous improvement over the idiots that normally operate cars): Turning Across Bicycle Lanes 21717. Whenever it is necessary for the driver of a motor vehicle to cross a bicycle lane that is adjacent to his lane of travel to make a turn, the driver shall drive the motor vehicle into t…

Interesting thing about that law is that it's probably there due to the limitations of human drivers. An autonomous vehicle that is fully aware of objects on every side probably doesn't need the precaution that the law requires. It will only turn when there is no cyclist coming.

Re: The latest chapter for the self-driving car: mastering city street driving

#362

Post by prawn »

aetherson wrote: post #328
TheEzEzz wrote: post #307

Earlier quoted context omitted.

Paying for the self-driving cars will likely be a cost savings by itself, since we won't need as many auto-cars as regular cars. Most cars are only used maybe 5-8% of the time, and sit unused otherwise. An auto-car service could probably get at least 50% usage out of a car, potentially much more. That means a 10x reduction in the number of cards needed.

For this to be true, we have to assume that, during rush hour, at least 90% of all cars are parked and unused. 90%? That's a hell of a lot, dude. It's certainly true that most cars spend most of their time being parked and unused. But they tend to do so during times of day when there is very low demand. The vast majority of cars are unused at 4am (and, indeed, at 2pm), certainly. But even if every car that was on the…

Surge pricing equivalents will motivate people to adjust/stagger their working hours. I already arrive early and leave late to avoid peak hour.

Pod-based transport would also make public transport easier and closer to a door-to-door experience.

Re: The latest chapter for the self-driving car: mastering city street driving

#363

Post by Zigurd »

lepht wrote: post #315
theg2 wrote: post #234

Earlier quoted context omitted.

I ride a bike and drive a car. And while doing both, I do my best to obey the laws and not injure other people. In Boston, and more specifically in my area of Cambridge we have major issues with drivers turning without looking at the bike lane and bikers running through red lights or riding on sidewalks. In MA, we have very specific laws regarding how bikes are to be treated under the law and whenever riding I abide…

I'm also in the Boston area (Somerville) and have been wondering about the proper way to respond to right turns with a cyclist in the bike lane behind me. When I have right-of-way and use my turn signal ahead of time, shouldn't I have the ability to turn right without waiting for the bike to pass (similar to how the cars behind me must wait for me to make my turn before continuing?)

I wait for the cyclist to stop or pass me.

Re: The latest chapter for the self-driving car: mastering city street driving

#364

Post by DavidAdams »

cbhl wrote: post #31
nabla9 wrote: post #18

Earlier quoted context omitted.

Somebody correct me if I'm wrong, but Google cars still can't drive in the rain or in the snow. Not even on a wet street. They have driven all those thousand miles in Mountain View, Calif. in good weather. Generally lidar's don't work well with rain or reflective surfaces.

My understanding is that it uses vision processing algorithms on a camera nowadays, and works just fine in rain. Snow is an issue because the car can't see the lane markings through the snow -- a human drives through snow by remembering where they were or making a guess.

There's a relatively simple fix for lane markings, that would help autonomous cars even in good weather. You know those reflective lane marking doodads they embed in the road surface? We start replacing them with magnetized nails that are pounded into the asphalt in intervals. Sensors in the car could recognize the lanes even in deep snow or pea-soup fog. And homeowners could easily outfit their driveways with a trip to Home Depot.

Re: The latest chapter for the self-driving car: mastering city street driving

#365

Post by intended »

encoderer wrote: post #272
chillingeffect wrote: post #256

Earlier quoted context omitted.

Way to only see everything with rose-colored glasses. Has it even occurred to you there might be any possible cons to your scenario? First of all, the lot the cars come from is going to require real estate, so subtract that from your windfall. Where will they be at night or other low-use times? Second, subtract the additional resources in gas/electricity it takes to get to my house from god-knows-where it came from.…

You sound like you live in the suburbs. I suspect car-as-a-service wouldn't be popular there. Honestly, if you live close enough to the store that you can drive there in under 5 minutes, to me that's a sign that you don't need to drive there. Where I live (Berkeley), and in the Sf Bay Area in general, car as a service would be remarkably effective and popular.

do people realize that all these space savings mean that builders will now make smaller houses and more of them?

Re: The latest chapter for the self-driving car: mastering city street driving

#366

Post by RogerL »

Qworg wrote: post #258
jessriedel wrote: post #194

Earlier quoted context omitted.

Since you're one of the few people here doing more than speculating, could you expand on this a bit? Is the LIDAR resolution a hard physical limit, and if so what's the origin? What makes identifying the debris with the camera so hard, just the standard object recognition problem in computer vision?

Certainly. LIDAR resolution is related to a few things: - How many beams you're putting out and at what angles - How fast you're spinning and how fast you can get the data back. - How many points of data you can get through your bus If the Google Car is using the big Velodyne ( http://velodynelidar.com/lidar/hdlproducts/hdl64e.aspx ), it puts out 1.3 Mpt/second over a 26.8 degree vertical field of view. That makes it…

But you are filtering the LIDAR data (source: I have I Velodyne LIDAR in my office). Yes, one stray hit will be ignored. But, you aren't getting one stray hit. You get 20 hits on this revolution, 6 on the next, 21 on the next, and so on, and then use various algorithms to determine if you are seeing objects or noise. Recall that besides the inaccuracy of the spinning lidar (it does not hit exactly the same spot even when sitting still and hitting an immobile object), the car is moving. Any decent Bayesian filter derives a heck of a lot of information from these changes. I'm not talking about the physical change, which is also important (change of position = change in angle = different reflection point). I'm talking about how your process model generates information via hidden variables - from position over time we derive velocity, and the correlation between velocity and position increases our accuracy. In the context of the LIDAR, the filter can detect that there is a small clump of points moving towards the car at 70 mph (from it's point of view, of course from our POV it is the car approaching the debris at 70 mph). Reflections that are true noise will not be clumped, and will not consistently have a velocity that matches the car's.

With all that said, I've just played with the LIDAR, I can't give you lower detection limits. But with proper algorithms it is better than the static computation suggests.

Re: The latest chapter for the self-driving car: mastering city street driving

#368

Post by intended »

Reading a lot of comments here, the theme seems to be that the benefits will trickle down to the average consumer. I can see immediate cases where it's just going to go to those with sufficient economic muscle.

Less need for a garage means smaller houses, and more of them (for example).

Improved insurance? Probably the same insurance premiums (inflation adjusted), with a higher premium for more human miles driven.

If there is money on the table, it will get taken by those with the wherewithal to grab it.

Re: The latest chapter for the self-driving car: mastering city street driving

#369

Post by Qworg »

RogerL wrote: post #366
Qworg wrote: post #258

Earlier quoted context omitted.

Certainly. LIDAR resolution is related to a few things: - How many beams you're putting out and at what angles - How fast you're spinning and how fast you can get the data back. - How many points of data you can get through your bus If the Google Car is using the big Velodyne ( http://velodynelidar.com/lidar/hdlproducts/hdl64e.aspx ), it puts out 1.3 Mpt/second over a 26.8 degree vertical field of view. That makes it…

But you are filtering the LIDAR data (source: I have I Velodyne LIDAR in my office). Yes, one stray hit will be ignored. But, you aren't getting one stray hit. You get 20 hits on this revolution, 6 on the next, 21 on the next, and so on, and then use various algorithms to determine if you are seeing objects or noise. Recall that besides the inaccuracy of the spinning lidar (it does not hit exactly the same spot even…

Of course. Filters do wonders - but they're still not a panacea.

You also get the covariance across that filter. Which means you still need to decide how much to cut, whether or not to reject as rain, dust or a reflection, whether to trust your reflections under a certain reflectivity value. Data helps with this, but, as always, there's a ton of corner cases.

Re: The latest chapter for the self-driving car: mastering city street driving

#370

Post by Qworg »

Schwolop wrote: post #338
Qworg wrote: post #285

Earlier quoted context omitted.

Well, LIDAR doesn't have that differentiation problem - it has a resolution problem for small (sub half meter) things at range. Stereo LIDAR would just add more points - you can correlate the two, but you'll still have resolution issues.

Well there are "Flash Lidar" sensors, that are effectively instantaneous depth cameras firing a lidar beam for every pixel - but a) they're hugely expensive, and b) they don't seem to have taken off (possibly because academia can't afford them to do the interesting research that'll help them take off)

There are some cheaper ones now - the K4W2 is a flash LIDAR.

But range is always difficult. I think PMD is in use in BMWs. They purport a 90 meter(!) range, but I've never played with one.

Post reply on HN