How are scientific research projects planned? In particular computer science, but possibly there exists processes for all research?

Possibly some context: Take any research endeavor. Finding a vaccine. Going to the moon. Clean energy. I don’t have a background in these scientific areas so picking one closer to home (Comp Sci) might be better. But the idea is how to battle the Unknowns? How to do it economically? How to make progress while not getting analysis paralysis?

Some might suggest a scrum, or an agile process try to solve these questions. I’m not certain that they address the same level or kinds of Unknowns.

Are there previous experiences that work, and those that don’t work? And why? The questions grows on the way forward through unknowns, and by definition the Unknown doesn’t exactly have a road map, and new context is developed regularly.

This maybe simply the question of the ‘meta’ variety: Is there research on comp sci research? If so does Comp Sci research have approaches or techniques that they rely on?