News for hackers. Stuff that matters.
Re:Uber’s self-driving car could not detect pedestrians outs... by chooseaname on Wednesday November 06, 2019 @08:43PM (#21469635)
> Classification is not needed to avoid objects. The key words here are actually "at each change, the object's tracking history is unavailable" and "without a tracking history, ADS predicts the bicycle's path as static" which is a horrible oversight.

Was it an oversight or were they having trouble with the algorithm(s) by keeping the object history and decided to purge it?

Re:Uber’s self-driving car could not detect pedestrians outs... by a_t48 on Wednesday November 06, 2019 @09:23PM (#21469821)
Pure speculation incoming: probably a completely unintended bug. Classify as static, static objects shouldn't have tracks, clear the history. Your object is reclassified as dynamic, whoops, we already cleared the history, oh well.
Re:Uber’s self-driving car could not detect pedestrians outs... by vilhelm_s on Wednesday November 06, 2019 @10:05PM (#21470038)
Static is not a classification; in this case the classification was cycling between "unknown", "vehicle", and "bicycle", and each time the classification changed the history was reset. Then each detected object can be either static or be assigned a predicted path, but unknown objects without a history are always treated as static.

In any case, the history clearing seems to be a bug, they updated the software to keep the position history even when changing the classification.

Re:Uber’s self-driving car could not detect pedestrians outs... by a_t48 on Thursday November 07, 2019 @01:17AM (#21470893)
In your example "unknown" is implicitly "static". Either way, neither of us have access to Uber's source history to say what exactly went wrong. They do claim to have fixed it though, you're right.
Re:Uber’s self-driving car could not detect pedestrians outs... by jobigoud on Thursday November 07, 2019 @03:27AM (#21471460)
Strange that the history itself is not used for the classification.