It's interesting work. However, I wonder what problem it is trying to solve.

For at least several decades (and probably much longer), people in the aviation field have discussed and thought about the problems that arise when operators must refocus their attention on flying when they need to take over from autopilot, particularly when they are responding to some external alert.

Things happen much more quickly in a car, and the operators typically are not trained as well. I don't see a good solution involving a human driver "taking over" in situations that arise suddenly, because no matter how much you do on the automation side, the human side is not going to improve.