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

Another Tesla on autopilot steers towards a barrier

reddit.com

Post reply on HN
Print view View original

371–380 of 481 posts

Re: Another Tesla on autopilot steers towards a barrier

#371

Post by MattConfluence »

binaryjay wrote: post #359
lisper wrote: post #122

Here is my armchair diagnosis: right before the car veers towards the barrier it drives through a stretch of road where the only visible lane marker is on the left. Then the right lane marker comes into view at about the point where the lane starts to widen out for the lane split. The lines that will become the right and left lane markers of the split left and right lanes respectively are right next to the van in fro…

>Here is my armchair diagnosis: Odd how your "armchair diagnosis" matches perfectly with the top ranked comment in the reddit thread that was posted 1 day ago. courtlandreOwner 815 points 1 day ago It sees the white line on the left, the white line on the right and thinks its a big lane. Its trying to center the car in the lane.

Is it really odd that two people could reach the same conclusion when presented with the same video?

Re: Another Tesla on autopilot steers towards a barrier

#372

Post by gizmo385 »

lsaferite wrote: post #335
niftich wrote: post #207

Earlier quoted context omitted.

Cruise control is great where there's long stretches of open road, light traffic (LOS A or B [1]), and little need for frequent braking. The star usecase is to set cruise to be very near the speed limit, such that after acceleration events like overtaking, you coast back to highway speed. It's a low-effort way of ensuring that one will be compliant with speed laws most of the time, yet maintaining a steady pace. I to…

> if it gets disengaged for any reason then you can easily overrun into the car ahead Are there cars that have ACC and don't have AEB (Automatic Emergency Braking)?

Furthermore, if an ACC system is getting disengaged, you're going to coast to a stop. Not accelerate straight into the car in front of you.

Re: Another Tesla on autopilot steers towards a barrier

#373

Post by ocdtrekkie »

syntaxing wrote: post #283

Does anyone know if all these recent accidents are with the old MobilEye technology or the new Nvidia system? Or was it a rumour that Tesla was moving towards Nvidia rather than MobilEye (I think MobileEye was the one that didn't want to renew their contract?)?

I would be careful calling it just the MobilEye system or the Nvidia system and assigning blame accordingly. They both create autonomous driving components, but at the end of the day, Tesla, Uber, etc. write their own software, and are responsible for the whole of the systems that are actually in their cars.

Re: Another Tesla on autopilot steers towards a barrier

#374

Post by clickok »

DSingularity wrote: post #182
dancek wrote: post #165

Earlier quoted context omitted.

> It's not possible to engineer an autonomous system that never fails, but it is possible to engineer one in such a way that it never fails to detect that it has failed. This is a very good point: just like a human driver should slow down if they can't observe the road ahead well enough, an AI should slow down when it's not confident enough of its surroundings. This is probably very difficult to do, and I'm skeptical…

I feel that’s the problem. Techniques like SVM can provide reasonable definitions of confidence. Reinforcement Learning.. DNN... the mechanisms behind autonomous cars... are they able to do this??

Yes.

In particular, it's possible to learn the variance of the return using TD-methods with the same computational complexity as learning the expected value (the value function). See [0] for how to do it via the squared TD-error, or [1] for how to estimate it via the second moment of the return, and my own notes (soon to be published and expanded for my thesis) here [2].

It turns out that identifying states with high variance is a great way of locating model error-- most of the real-world environments are fairly deterministic, so states with high variance tend to be "aliased" combinations of different states with wildly different outcomes. You can use this to improve your agent via either allocating more representation power to those states to differentiate between very similar ones, or have your agent account for variance when choosing its policy. For example, Tesla could identify when variance spikes in its model and trigger an alert to the user that they may need to take over.

Additionally, there's work by Bellemare [3] for estimating the distribution of the return, which allows for all sorts of statistical techniques for quantifying confidence, risk, or uncertainty.

---

0. https://arxiv.org/abs/1801.08287

1. https://arxiv.org/abs/1607.00446

2. http://rl.ai/posts/fun-with-the-td-error-part-ii.html

3. https://arxiv.org/abs/1707.06887

Re: Another Tesla on autopilot steers towards a barrier

#375

Post by Piskvorrr »

jboles wrote: post #307
michaelt wrote: post #82

Earlier quoted context omitted.

Guess what computers excel at? Driving consistently on consistent highways. Watching the video, would you say the computer was excelling, or that the road was radically unusual? I wouldn't.

The location is the end of some tidal flow lanes rejoining the main freeway lanes with an unusual configuration of lane markings and barriers. I drive past the spot all the time and can totally see a computer messing up here.

In other words, "works great if it works, might actively try to kill you at random, there's no clear way of telling when it might try the latter, it's your fault in any case." Yeah, that's very much the safe driving paradise I'm being promised "in two years" for a decade now, by the technooptimists. It's almost as if the technology is at 50%-ready instead of the 90% that marketing types seem to peddle, who would have thunk?

Re: Another Tesla on autopilot steers towards a barrier

#376

Post by Piskvorrr »

xtrapolate wrote: post #116
vorpalhex wrote: post #63

Earlier quoted context omitted.

So human drivers are terrible at driving for any long period of time, especially on relatively consistent "boring" highways. Guess what computers excel at? Driving consistently on consistent highways. The Tesla Autopilot is supposed to be the always aware and paying attention portion on these cases where a human driver would be very likely to start texting or dozing off. Now, it's not fully autonomous and may well de…

> "Guess what computers excel at? Driving consistently on consistent highways." Possibly on paper. In reality, as of right now, computers are clearly far from excelling at this specific task.

Assuming perfectly spherical cars on perfectly consistent highways in a vacuum.

Re: Another Tesla on autopilot steers towards a barrier

#377

Post by return0 »

bennettfeely wrote: post #332
return0 wrote: post #36

AV driving seems to be ripe for disruption

Any 16 year old can already "disrupt" the industry just by getting a license and driving. Humans are and will continue to be far more capable, rational, and economical drivers for the foreseeable future. Sounds similar to tech bros who want to radically "disrupt" carbon sequestration. Here's an actually proven idea: plant a tree.

maybe it was not obvious , AV = autonomous vehicles

Re: Another Tesla on autopilot steers towards a barrier

#378

Post by greedo »

ucaetano wrote: post #300
hw wrote: post #297

Earlier quoted context omitted.

I think more people need to embrace and understand the fact that we are a long way from having fully autonomous vehicles that you can just give the steering reins to from start to finish. The term 'Autopilot' gives off the wrong idea that the driver can sit back and relax while the car brings you from point A to point B safely. The same is happening with chatbots - more and more businesses think they can put a chatbo…

> I think more people need to embrace and understand the fact that we are a long way from having fully autonomous vehicles that you can just give the steering reins to from start to finish Apparently, we're not, as Waymo has shown. > The term 'Autopilot' gives off the wrong idea that the driver can sit back and relax while the car brings you from point A to point B safely. Agreed, the "but that's not how autopilots w…

Waymo has shown no such thing. In controlled, limited circumstances, yes, but in complex, poorly mapped, degraded conditions, not yet.

Re: Another Tesla on autopilot steers towards a barrier

#379

Post by LeoPanthera »

singingboyo wrote: post #255
Someone1234 wrote: post #246

Earlier quoted context omitted.

Subaru's system is simply fantastic and you can tell they thought a lot about the "human condition" (human weaknesses) while designing it. Lame keep assist is very subtle, I describe it like wind blowing on the side of the car. You can trivially over-power it and it will beep if you exit the lane (without blinker on) with or without lane keep assist active. The whole package of safety features is wonderful and impres…

I drive a Nissan myself, but from the sounds of it it's about the same. LDW (just warning in my case), FEB + pedestrian/moving object detection, blind spot detection, intelligent/adaptive cruise control. (There's also a lane keeping system option called ProPilot, but I skipped out on that - I don't think I'd trust it in the regular rain anyway) Anyway, my point is these things are common now, and while getting the fu…

I have ProPilot. It's pretty great. In heavy rain it shuts itself off and tells you that it is unavailable in bad weather.

Re: Another Tesla on autopilot steers towards a barrier

#380

Post by Piskvorrr »

jerrycruncher wrote: post #295
maxerickson wrote: post #84

Earlier quoted context omitted.

It's not any different than the situation with bad human drivers, but anyone on the same road is exposed to these systems.

It is different, though. You can't push an update to stop humans from endangering themselves and/or others. Tesla, however, is able to do exactly that with their cars. That they haven't, when their technology has been proven to be fatally dangerous, is unconscionable.

And conversely, you can't push an update that causes humans to start endangering themselves. (Unless Tesla has some magic way of preventing regressions, which doesn't seem to be the case)
Post reply on HN