This paper is a core part and fundamental research evolving into how C++ could eliminate UB from the language in the future via a hybrid approach: subsetting, runtime checks, annotations, new features.

I think it is worth a look.