Live data from Hacker News

Tesla auto-pilot keeps confusing moon with traffic light then slowing down

reddit.com

21–30 of 93 posts

Re: Tesla auto-pilot keeps confusing moon with traffic light then slowing down

#22
post #5
post #3

I only have two cameras and have never confused the moon with traffic lights.

If your head was in a vice and your eyeballs were superglued into the straight-ahead position you might have more trouble.

That's not a great comparison, since our focal area is different than that of cameras. I'm pretty sure anyone looking at the camera feeds in a Tesla wouldn't be confused either; this is a software problem.

Re: Tesla auto-pilot keeps confusing moon with traffic light then slowing down

#24
post #18

This is a problem with learning the appropriate context. In the case of a yellow traffic light, there should be a traffic light fixture and two other (unlit) lights. Without this context, it becomes apparent the yellow pixels in a circular shape (representation to the network) are not a traffic light at all. Tesla’s vision/ml systems are amazing. I would love to learn more about how unit testing for this type of erro…

But in foggy or dark situations, a yellow light can indeed appear as a floating yellow circle.

Re: Tesla auto-pilot keeps confusing moon with traffic light then slowing down

#26
post #16

I can't understand why instead of using all the possible senses: vision, auditory, radar, lidar they want to only use vision as if there's some sense of "purity" in doing that.

Which of the above would solve this problem?

Re: Tesla auto-pilot keeps confusing moon with traffic light then slowing down

#27
post #18

This is a problem with learning the appropriate context. In the case of a yellow traffic light, there should be a traffic light fixture and two other (unlit) lights. Without this context, it becomes apparent the yellow pixels in a circular shape (representation to the network) are not a traffic light at all. Tesla’s vision/ml systems are amazing. I would love to learn more about how unit testing for this type of erro…

> Tesla’s vision/ml systems are amazing

More importantly, they're the only ones accessible to the average consumer. Waymo/Zoox/Daimler all have equally if not more impressive systems.

A real issue with tesla is that they want to be vision-only, which is going to make getting to level 5 first almost impossible.

BTW - knowing where the moon is happens to be an extremely solved problem.

Re: Tesla auto-pilot keeps confusing moon with traffic light then slowing down

#28
post #2

Use lidar, goddamnit.

I wonder if the concern is that LiDAR looks "ugly" or nerdy on the front of the car. And a bit more aero drag.

Elon can't afford to put LIDAR on cars and sell them to a large audience. It's still too expensive. So instead, he makes excuses for why they don't do it while rolling out objectively worse perception systems.

Re: Tesla auto-pilot keeps confusing moon with traffic light then slowing down

#29
post #2

Use lidar, goddamnit.

I wonder if the concern is that LiDAR looks "ugly" or nerdy on the front of the car. And a bit more aero drag.

Lidar can be embedded in the headlights or on the rear view mirror; there is no need for ugly protrusions.

Re: Tesla auto-pilot keeps confusing moon with traffic light then slowing down

#30
post #18

This is a problem with learning the appropriate context. In the case of a yellow traffic light, there should be a traffic light fixture and two other (unlit) lights. Without this context, it becomes apparent the yellow pixels in a circular shape (representation to the network) are not a traffic light at all. Tesla’s vision/ml systems are amazing. I would love to learn more about how unit testing for this type of erro…

You can learn about their systems by watching talks by Andrej Karpathy. As a robotics engineer interested in vision, their architecture is inspiring. This talk [1] is a good overview but each talk he gives is a little different so search for more if you want to know as much as possible.

But the big thing is that their autonomy computer can be programmed to look for odd scenarios and send them back home. Tesla uses their fleet of hundreds of thousands of cars to collect edge cases like this, and then they have a kind of compartmentalized neural network system that breaks apart disparate tasks. With their collected examples they can create unit tests to ensure that the moon stops activating the stoplight detector. Once trained, the unit tests presumably help ensure they don't end up with future regressions.

So basically every time you see a Tesla do a weird thing, there is a good chance it will stop doing it soon enough. At least if it's hitting hacker news.

[1] https://www.youtube.com/watch?v=hx7BXih7zx8

Post reply on HN