Re: Kalman Filtering
DAM admitted, "I dunno..."
I have failed to make myself sufficiently clear. As Fermat noted, these damn margins are too small...
Kalman filtering is an optimum method of combining noisy data inputs to achieve a unified solution. Aircraft use it to mix GPS, Inertial and other data to provide a single navigation solution that is reliable.
The point is in response to the following issue.
Slightly expanded extract [this issue->]: "...'localization errors', where... ...the vehicle was unable to really work out its position in relation to its environment... Errors included... ...not braking hard enough for a stop sign."
Clearly the designers of autonomous vehicles can't even get the basics correct. .: They're utter morons.
The mention of the Apollo computer was intended to convey that Kalman filtering is an old and well established technique, and doesn't require GBs of software.
Kalman filtering is non-trivial, but it's the sort of thing that should have been included on the Requirements spec from Day 1.
These people are clowns. Governments need to get a big bucket of Roadworthiness Regulations and dump it on them.
The AV industry clearly needs some Adult Supervision.