3 ms·

This seems very hard for Tesla to fix with a simple software patch. If the car is using cameras to look at the lines on the road + radar which can not see stati

This seems very hard for Tesla to fix with a simple software patch. If the car is using cameras to look at the lines on the road + radar which can not see stationary objects well, then I don't see how they'd have enough sensor information to fix this reliably.

ealexhudson 9y ago
In principle the cameras can see the stationary object like we can, after all, we don't have LIDAR. Fundamentally, though, that does seem like the long route to self driving. Being able to not drive into the proverbial wall seems like Job #1 of a pilot system. Tesla need to be careful they're not not creating a reputation for themselves at this point.
simion314 9y ago
They have many cars passing those places, can't they improve the maps they have, in the places where the video inputs and Tesla maps will conflict decide to slow down and ask the human to take over, then analyze the issue and add is as a unit test.
- Piskvorrr 9y ago
  Improving maps only papers over the issue at this one particular place. What happens when this happens somewhere else, where there's no marking in the map instructing the lane-keeping software "this is a known bug, follow the right lane marker here"?
  - simion314 9y ago
    I don't mean of marking this spot, but soemthing like 1 this is a new road, "autopilot" will not engage but collect data and add it to the map 2 after 100 or other better number of cars passed that point and the maps are updated enough AP can engage 3 if radar+camera detects something that does not match the maps it slows down and asks the human to take over, the maps are used as a way to double check what the AI wants to do. This system will not work on small roads with few Tesla cars on them but it will prevent problems on big roads and collect useful data to add to the unit tests
    - Piskvorrr 9y ago
      In other words, better roads get better, worse roads get nothing. Isn't that reinforcing the current problem of "SDVs drive well in areas where SDVs drive well; beyond the map edge there be dragons"? See my reply earlier: https://news.ycombinator.com/item?id=16706856 https://news.ycombinator.com/item?id=16706856
      - simion314 9y ago
        Is not this better then no road gets safer? At least high traffic roads get safer, less damages and live lost, fixing the bugs that are discovered here could improve the situation on smaller roads. My idea was how they could act on this situations with the drive assist tech(stupidly named autopilot) I am not of the opinion that true self driving cars are around the corner, not even in major US cities.
        Piskvorrr 9y ago
        Sure, for autonomy ~level 3, this would be okay. I, too, dislike the rhetoric "just patch this one specific killing bug and voila, full auto!"

ealexhudson 9y ago

In principle the cameras can see the stationary object like we can, after all, we don't have LIDAR. Fundamentally, though, that does seem like the long route to self driving. Being able to not drive into the proverbial wall seems like Job #1 of a pilot system. Tesla need to be careful they're not not creating a reputation for themselves at this point.

simion314 9y ago

They have many cars passing those places, can't they improve the maps they have, in the places where the video inputs and Tesla maps will conflict decide to slow down and ask the human to take over, then analyze the issue and add is as a unit test.

Piskvorrr 9y ago
Improving maps only papers over the issue at this one particular place. What happens when this happens somewhere else, where there's no marking in the map instructing the lane-keeping software "this is a known bug, follow the right lane marker here"?
- simion314 9y ago
  I don't mean of marking this spot, but soemthing like 1 this is a new road, "autopilot" will not engage but collect data and add it to the map 2 after 100 or other better number of cars passed that point and the maps are updated enough AP can engage 3 if radar+camera detects something that does not match the maps it slows down and asks the human to take over, the maps are used as a way to double check what the AI wants to do. This system will not work on small roads with few Tesla cars on them but it will prevent problems on big roads and collect useful data to add to the unit tests
  - Piskvorrr 9y ago
    In other words, better roads get better, worse roads get nothing. Isn't that reinforcing the current problem of "SDVs drive well in areas where SDVs drive well; beyond the map edge there be dragons"? See my reply earlier: https://news.ycombinator.com/item?id=16706856 https://news.ycombinator.com/item?id=16706856
    - simion314 9y ago
      Is not this better then no road gets safer? At least high traffic roads get safer, less damages and live lost, fixing the bugs that are discovered here could improve the situation on smaller roads. My idea was how they could act on this situations with the drive assist tech(stupidly named autopilot) I am not of the opinion that true self driving cars are around the corner, not even in major US cities.
      - Piskvorrr 9y ago
        Sure, for autonomy ~level 3, this would be okay. I, too, dislike the rhetoric "just patch this one specific killing bug and voila, full auto!"