AlanSubie4Life
Efficiency Obsessed Member
Well you are thinking like a human, where it is relatively trivial for you to identify and respond to those objects same as if they were cars (putting aside for the scenario of objects falling off a lead vehicle, Mercedes wouldn't be at fault). But we've seen clearly this is not the case for most ADAS. For most systems, they have long been rock solid in responding to a lead car, even with relatively simple sensors (that has long been how ACC was designed), but responding to objects that are not cars is decidedly a mixed bag. It's not that the car is unable to do that at all, it's that it helps in the probabilities, as even the best L4 solutions today are not flawless (for example even Waymo still has random halting according to latest reports, even though it is very rare).
One thing to think about is that the lead vehicles for the AC unit didn’t really slow down that much. They just went around it. We were going 80mph and maybe got down to 55mph at the lowest - it is dangerous to slow down on the freeway and traffic conditions were such that slowing a lot did not make sense since you could just safely swerve into adjacent lanes (it was in #2, slightly toward the #1, and I went for #2.5).
Obviously a bit off topic at this point since the Mercedes would be traveling much more slowly. Just pointing out that there are many situations where obstacles are avoided without slowing - there is no reason for a lead vehicle without lead vehicles to slow since it can see the object coming for hundreds of yards.
I think the video had an example of a “cut out” though. Hard to tell how well it was handled exactly without accelerometer data.
I wonder what the minimum allowed distance is? I’d like to see a cutout for the surprise reveal of a stationary vehicle at maximum speed, last second, at minimum following distance. Not sure how comfortable it would be. Obviously they would have tested this.
Last edited: