Welcome to Tesla Motors Club
Discuss Tesla's Model S, Model 3, Model X, Model Y, Cybertruck, Roadster and More.
Register

The Real Use for Auto-Pilot

This site may earn commission on affiliate links.
Completely agree with AP 2.5 -- Even with Chill mode turned on. My sense is that the forward cameras aren't looking far enough ahead to see that things are slow. I think it comes down to a couple of things:
1. not looking far enough ahead to see a car in my lane that I can see as the driver.
2. Even if it DOES see the car up ahead, if the car is stopped or moving very slowly, it takes too long to figure out that the car is stopped/slow (we know there are issues with detecting "stopped" vehicles) and so it happily continues accelerating to the set speed. Then figures out that the car ahead is stopped or slow and slams on the brakes.

I haven't figured out yet WHY it cannot figure out a stopped/slow vehicle. This is one of the most basic things that our personal neural nets are very good at. Just from 101 traffic in the Bay Area there should be more than enough data to train the NN to detect a stopped vehicle in a traffic lane.

There's probably a test version of the software out there (or was) that stops great for stopped vehicles in your lane, with the downside that it also brakes hard at every "stopped" roadsign on the side of the road, or "stopped" traffic divider, or "stopped" overpass, etc. :)

I think it sometimes reacts to the car in front of the car in front of you slowing down. I'm also not sure how far ahead it actually looks, since I can see it can slow down based on cars slowing that don't even appear in the visualization, so it sees beyond that visualization for sure.