241–250 of 445 posts
Earlier quoted context omitted.
As it should, what a ridiculous thing to do.
Says the person who has never seen the ridiculous thing that is single lane road deadlock.
I'm looking forward to seeing the full NTSB report. This hints that Uber's mindset is focused on "obstacles", not "flat road". The first job of automated driving is to drive only on flat surfaces. Doesn't matter why it's not flat. Then you worry about where to go. That's how the off-road DARPA Grand Challenge vehicles had to work. "Flat road" is a pure geometry thing. Not much AI needed. On road, non-flat road is rar…
"flat road" => assume you can't continue until you prove it's safe
"obstacle detection" => assume you can continue unless there is something blocking the current path
The bit of autonomous vehicle work that I did started from a set of potential trajectories that were then pruned by obstacle detection. So we always assumed you could continue moving forward unless every potential trajectory was pruned. We ignored the cliff problem as beyond our scope.
Earlier quoted context omitted.
Said cynic sounds pitchforky. Even HIPAA tries to work with non-compliant companies to get them compliant before resorting to dishing out their notoriously expensive fines. At the end of the day, the NTSB isn't in the game of shutting down companies. NTSB would much rather see that self-driving vehicle development is happening safely since the matter of the fact is that it _is_ happening, despite all of its imperfect…
Hiring a former NTSB chair while there's an active NTSB investigations sounds a lot like a revolving door to me.
One does not just hire this guy as if he was some sort of money-chasing sleazeball. I mean, just google the guy's bio
Earlier quoted context omitted.
That, and the inattentive safety driver. It's quite infuriating really, letting lose a ton and a half of fast-moving mass without even the decency of looking where it's going. There's really no excuse.
yes for sure. I wish articles about this crash would continue to hammer on the basics, to keep front¢er this crash was completely avoidable. Even if there were no technical issues, (1) the driver was on her phone before and during the crash. The whole reason for the driver to be present is to intervene in a situation like this. (2) the car was driving too fast for conditions. Nobody should drive this fast in the…
Earlier quoted context omitted.
> so I don't see how doctoring the dashcam video would fit into the narrative. Well, at the end of the day this is all involves a strong element of PR & perceptions, so regardless of the quality of other data, it would be very much in Uber's interest to have a super dark video be in the initial stories about this accident.
> it would be very much in Uber's interest to have a super dark video Why? As soon as the video came out, the first thing people (correctly) pointed out was that the dynamic range of the dashcam was crap. Besides, everyone knows that self-driving cars are essentially giant self-telemetry collection machines, so one would have to be either extremely short-sighted or incredibly stupid to think they would be able to get…
Here is the more likely order of events:
1. Driver calls supervisor
2. Supervisor redirects call to senior Uber lawyer
3. Uber lawyer says "Don't do anything until we get to the scene. Be polite to police but tell them to wait until our legal team arrives"
4. Legal team at scene, talks with uber CEO (and senior lawyers over phone) then after an hour or so tells NTSB they will hand over footage after engineer arrives and extracts the dashcam footage
5. Engineer hands dashcam footage to lawyers, who then hand it to NTSB
Earlier quoted context omitted.
Don't protest self-driving cars, protest the laws you feel shouldn't apply to you. After all, there are plenty of other reasons jaywalking laws might become more strictly enforced.
The term "jaywalking" was a slur invented by car companies to steal rights from pedestrians. If cars can't avoid people, the cars need to be separated by physical barriers or removed.
A pedestrian may illegally have entered a road, but still have right-of-way, and self driving cars will need to account for that.
However, protesting self driving cars will not change whether walking in the street is legal, since it already is not.
Therefore, whether you do so and get fined by police is entirely up to their discretion, which could change at any time for any reason.
If you want it to be legal, protest the illegality, not the car company.
The exact circumstances of the struck pedestrian likely violated many [possibly curde] bayesian priors for what to consider a positive collision threat: A rogue pedestrian at an unflagged portion of road, at an odd time of night, on a road that doesn't usually have pedestrians, with a bike moving perpendicular across the road (instead of parallel along it). Each one of these had been learned to be a high false positi…
In many of the cities where Uber would like to deploy self-driving cars, a "rogue pedestrian" crossing where there is no marked crosswalk would be at least a multiple-times-per-day event. Any system trained to treat that as an ultra-rare anomaly would be a killing machine if unleashed in such a city.
I want (and expect the system already does) account for this difference. And I also suspect the model to apply this pedestrian expectation modelling is imperfect and could use improvement. I don't know how to make it perfect, or if that's even a reasonable goal.
Earlier quoted context omitted.
No, what you're describing sounds a hell of a lot safer. (It's also how people make left turns in Vancouver. The fact that it's illegal in Seattle infuriates me to no end.) The Pittsburgh left sounds like a recipe to get T-boned by an inattentive driver. Pulling into the intersection, and making a left turn when it is safe is a much better alternative.
It is legal to pull into the intersection on green or yellow, and wait until it is clear (often when the light turns red) to turn left in Seattle. It is the correct/legal way to make a left turn. Otherwise you are just going to sit there blocking traffic for multiple cycles during busy times.
Earlier quoted context omitted.
Hiring a former NTSB chair while there's an active NTSB investigations sounds a lot like a revolving door to me.
Um, to go back to being NTSB chair, he would have to be re-appointed by the POTUS... One does not just hire this guy as if he was some sort of money-chasing sleazeball. I mean, just google the guy's bio
You are not hiring them for their bio, achievements or abilities. You are hiring them to signal existing and future regulators that being nice to you can be highly profitable.
The exact circumstances of the struck pedestrian likely violated many [possibly curde] bayesian priors for what to consider a positive collision threat: A rogue pedestrian at an unflagged portion of road, at an odd time of night, on a road that doesn't usually have pedestrians, with a bike moving perpendicular across the road (instead of parallel along it). Each one of these had been learned to be a high false positi…
I am rather confused: Should the car always respond by slow down as the mandatary response when there is object in front? Are you implying that the algorithm would conclude that because it's unlikely that there is an object in that circumstance, it is ok to disapprove its own observation that there is some thing in front of the car? That sounds utterly nonsensical.
If there's truly no possible route to avoid some collision and therefore the car needs to decide which collision is preferred, then the statistics can kick in, but in this Uber scenario, it shouldn't have gotten that far. From the outside, it seems like a fundamental design issue.