The idea of using algorithms should be applied to the speed limit itself.
The concept of fixed limits is outdated given modern tech and borderline ludicrous on most roads.
An empty freeway that could safely permit speeds of 100 mph (in a modern vehicle,
with an experienced driver), whereas the exact same stretch of road might need a
limit of 25 mph on a snowy day or in a rainstorm (40), fog (20-50) etc.
Why not tie the "limit" to realistic parameters and then ding anyone breaking the variable speed ceiling displayed by the vehicle (or linked to the cruise control)?
I have often seen drivers on highways going at dangerous speeds in awful conditions,
but nonetheless technically "legal"; we need to drop this one speed fits all circumstances bs.
The other idiocy built into this study is blind and literal interpretation of law.
That may be convenient for the profit seeking highway robbers (you know what I mean),
but even in America that level of ass-hattery is fairly rare (but mindless legal stuff is happening much too often)
It's really disturbing to see the massive disconnect between this kind of academic study and any kind of reality.
Ok... academic faculty dweebs, here are some real world algorithms you should go figure out:
a. How to implement speed algorithms for safe but variable limits
b A study that shows the de-facto algorithm in use for cop pay and grade review cycles as related to ticketing stats (gotta be a function in there somewheres).
c. Develop for the FBI an algorithm that alerts the district attorney of "municipal highway robbery" scams, on the vast stretches of "permanent construction" zones, where no one ever really works but lots of tickets happen near monthly quota time...