Robotics/control theory has a systemic nomenclature problem
It seams that every gander towards building a more capable OODA paradigm reinvents the fundamentals only to achieve a marginally greater escape velocity of intuitional abstraction than the previous generation
Well the theory probably configures its own boundaries. There's been a lot of new machine learning in the last 20 years but this relegates it to a small corner etc. Someone else might make ML a second continent.