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

Waymo now testing its self-driving cars on public roads with no one at the wheel

techcrunch.com

Post reply on HN
Print view View original

491–500 of 634 posts

Re: Waymo now testing its self-driving cars on public roads with no one at the wheel

#491

Post by Johnny555 »

dagw wrote: post #490
Johnny555 wrote: post #471

Earlier quoted context omitted.

No, you can't. A rail line can carry many more passengers/hour than cars, even self driving cars. From https://letsgola.wordpress.com/2014/08/24/capacity-101/ A typical freeway lane today can carry around 3600 passengers per hour. Reduce the headway between cars to 1 second and put 4 passengers in every car and that goes up to 14,400 pax/hr However, run a heavy rail line and run trains every 3 minutes, and it can car…

The number of areas in the world where you have two points where it's worth running a heavy line train every 3 minutes is minuscule.

But, not coincidentally, those areas are the same areas where freeways are at capacity and would benefit from running 4 passenger self-driving cars at 1 second headways.

Self driving cars work well with railroads -- the cars can get passengers from their lower density neighborhoods to the rail stations, and rail can take the passengers to the urban centers where their jobs are. It makes little sense for a self-driving car to take a worker from his suburban home all the way to the crowded urban center where his car either needs to find someplace to park, or join thousands of other cars looking for someone else to ride until it's time to go home.

Re: Waymo now testing its self-driving cars on public roads with no one at the wheel

#492

Post by jasonkester »

bmsleight_ wrote: post #276
aidenn0 wrote: post #232

Earlier quoted context omitted.

I still think that proper self-driving cars are such a paradigm shift that predictions of how it will work are still up in the air. For example, right now I live on the west coast and my parents live on the east coast. 4 kids, Minimum of 40 hour drive or $3500 flight. Here's one tiny example: I can work remotely for a few weeks and my boss doesn't mind. The kids sleep for 8+ hours at night. Friday, Saturday, Sunday,…

Camper-van, pick-me up at the office (London) on Friday evening, by Saturday morning we in the highlands of Scotland. It sounds silly - but the more automation the more I want to spend within my fictional camper-van. Soccer away match - no problem. Trip to the beach again, overnight. It is like an sleeper train - only more comfortable and spacious, as it would be my families camper-van.

I get the impression that people who make this sort of comment have never ridden a night bus with beds.

It is in no way comparable to a sleeper train, which rides on relatively smooth, straight rails and isn't constantly swerving about and dumping everybody on to the ground. You need to "sleep" in this tensed up position where you're wedged in between ends of the bed using your core to keep you in place.

On the sleeper buses in China, you'll often see people standing in the aisle, getting a rest from "sleeping" while building strength for another bout of it.

So yeah, you'll get to Scotland the next morning. But next time you'll take the train and rent your van at the station.

Re: Waymo now testing its self-driving cars on public roads with no one at the wheel

#493

Post by gambiting »

dheera wrote: post #444
seanp2k2 wrote: post #418

Earlier quoted context omitted.

once people are okay with sharing seats in a van That's going to be a long wait in the US! People would rather pay 10x to sit in their own luxury bubble in traffic and have everyone see how fancy their luxury bubble is vs. /gasp/ sharing it with others. Most people here don't want mass transit, they want their big truck / SUV to haul 1-4 people ("it's safer for my kid").

> People would rather pay 10x to sit in their own luxury bubble in traffic I think it has a lot less to do with the bubble factor than the fact that public transportation in the US just sucks, period. My 3-mile commute from Menlo Park to Palo Alto would take - bus: 58 minutes (2 minute walk from home to bus stop + 12 minutes average wait for ECR bus + 10 minutes on ECR + 12 minutes average wait in Palo Alto Transit C…

I live in a UK city with fantastic public transport options, and I would get to work faster by taking public transport than by car about 90% of the time. And yet I still drive, because I just don't like sharing the space with strangers, I can sit in comfort at the temperature that I like and listen to music/podcasts that I want without using headphones. I'm pretty sure I'm not alone in this.

Re: Waymo now testing its self-driving cars on public roads with no one at the wheel

#494

Post by ekianjo »

Johnny555 wrote: post #471
ekianjo wrote: post #379

Earlier quoted context omitted.

You can have way more passagers on a highway than on railroads though. Capacity wise cars are the better choice because its a continuous flow. Plus, they get you where you need to be, not just where the track stops.

No, you can't. A rail line can carry many more passengers/hour than cars, even self driving cars. From https://letsgola.wordpress.com/2014/08/24/capacity-101/ A typical freeway lane today can carry around 3600 passengers per hour. Reduce the headway between cars to 1 second and put 4 passengers in every car and that goes up to 14,400 pax/hr However, run a heavy rail line and run trains every 3 minutes, and it can car…

Even in Japan where trains are tightly operated, you don't get trains every 3 minutes (talking about long distance trains, not metros). At best one train every 15 minutes on average between peak hours and more calm periods. That divides your capacity by 5 already, so you are now at around 10 000 passengers per hour with your trains.

Add to that that freeways don't have a single lane, but usually at least 2, if not 3, according to your numbers you get 7000~10 000 passengers per hour with cars.

Now consider than most of the trains are not running full, and don't run at night at all between 10 pm and 5 am, and you can clearly see that there is an inherent advantage in terms of flows towards cars (since traffic never stops) per day. Plus you have mass transportation on freeways with buses, too.

And still the major advantage of the car is that you can stop exactly where you want to go. The train may get you from A to B, but most people want to go to C, D, E or F and not just B so trains are always a trade-off unless you have a very standard route every single time.

And I'm not even talking about cost yet. For passengers, trains are way more expensive (even in Japan, where trains are highly developed) than the same trips by car (even though the freeways are heavily taxed).

Re: Waymo now testing its self-driving cars on public roads with no one at the wheel

#495

Post by wiz21c »

randycupertino wrote: post #482
landryraccoon wrote: post #332

Earlier quoted context omitted.

> Car usage will go up by an order of magnitude because a major cost of driving isn't the dollars, but the time. This is why I think self driving cars will make traffic worse, not better. The more comfortable it is to be inside a car and the less people mind long car rides, the more people will drive (self driving or otherwise). If road infrastructure is the same, that necessarily means a much higher density of cars…

> This is why I think self driving cars will make traffic worse, not better. Yeah, of course self driving cars are going to make traffic way worse! Who even thinks otherwise? Suddenly you are going to have millions of disabled and homebound seniors wanting the cars to take them to do grocery shopping and errands, kids being driven around to and from from school and practice, and low income residents sleeping in their…

another climate apocalypse... 'cos even if all of these cars are electrical, they still need quite a lot of energy to be built and, after a while, destroyed...

Re: Waymo now testing its self-driving cars on public roads with no one at the wheel

#496

Post by kajecounterhack »

emerongi wrote: post #253

Why are the videos always done in perfect conditions? I like drive.ai's approach a lot more where they demonstrated the capabilities at night and in heavy rain. It's much more impressive.

What looks impressive in a video is often smoke and mirrors. Just because your car is working in inclement weather for the duration of a video doesn't mean it's safe or that the behavior scales to all situations. Proving that it doesn't cause catastrophic problems is even harder (you need sim).

FWIW most autonomous cars work great at night. As for weather, nobody has heavy rain completely solved (yet) (afaik). It messes with both camera and lasers.

Re: Waymo now testing its self-driving cars on public roads with no one at the wheel

#497

Post by kajecounterhack »

dx034 wrote: post #484
dabeeeenster wrote: post #70

There's a positive feedback loop that will happen: As they get more cars on to the road, the log more real world miles, which makes the cars better, which gets more cars onto the road which log more miles... I wonder how valuable this is?

Not sure how much that is worth. Tesla collects a huge amount of data but that doesn't help as much as some had hoped. You need to interpret all that data which is expensive and complicated.

Yep! Some data points:

1. Storing large amounts of logs data is challenging and expensive for anyone. So even if you can collect data, good luck storing it all. Think about what a self driving car has to store: images, audio, 3d point clouds, all organized by time, ideally labeled in some manner.

2. Unorganized data is not useful. For example, part of the autonomous vehicle problem is proving that the car is safe for . You need simulation to ensure model changes don't incur regressions. I've heard rumors that Tesla didn't have a sim team as recently as 1.5 years ago -- a lot of people argue "you can't sim everything" but most mature players now have realized that sim is vital.

If your data is not organized and accessible, or you can't afford to store the data, your development velocity is going to be really bad and/or your safety characteristics will not be provable.

One argument I've heard is "well with each new model launch why not just run the model in ghost mode for all live cars on the road, and see if it causes any issues" -- many life-threatening situations are rare combinations of factors. If you don't record them, you can't count on them to happen again during your regression test period. And again, even if it worked, the time it takes you to run a regression test is now a lot longer than a team that has sim that can run O(hours-days).

Re: Waymo now testing its self-driving cars on public roads with no one at the wheel

#498

Post by skbohra123 »

a9a wrote: post #148

It will be interesting to see how Lyft's partnership with Waymo evolves given this tech announcement and Waymo's acknowledgment that they plan to offer their own ridesharing service. Lyft's approach to self-driving partnerships in general seems to rest on the assumption that self-driving providers will catch up to each other before any of them can fully handle everything a human driver can do. If this assumption is c…

You are just thinking too much!

Re: Waymo now testing its self-driving cars on public roads with no one at the wheel

#499

Post by giosch »

dx034 wrote: post #479
eggpy wrote: post #86

Earlier quoted context omitted.

I wasn't aware of these, but with anecdotes like these then it makes perfect sense to have available remote human overrides! I don't mean to insinuate that maps is perfect, just that the number of times the cars will find themselves in this specific situation is probably minimal.

Also local human overrides. I as a passenger want to determine the route to take. Maybe I just don't like the recommended route, maybe I know that it won't work. There's no reason why the passenger shouldn't be able to determine the route.

Only if you own the car. If you are using uber you cannot choose the route...

Re: Waymo now testing its self-driving cars on public roads with no one at the wheel

#500

Post by oblio »

dagw wrote: post #486
oblio wrote: post #483

Earlier quoted context omitted.

Are cars comparatively cheaper than in the Model T era?

I think a Model T was about $22k-$25k in today's money.

So parent's logic, as I was expecting, doesn't necessarily apply. We just started wanting more from our cars, we didn't necessarily make them cheaper. There are very cheap cars but most would consider them death-traps...

I imagine a self-driving car packs a ton of servers, I'm not sure it can be cheaper than your average mid-range car...

Post reply on HN