Yes—this problem is generally known as bulb ramping (or “bramping”) for day-to-night timelapse. Existing commercial tools and open-source projects do this by gradually adjusting exposure to keep transitions smooth.

The hard part is that exposure control is partly predictive: sunrise and sunset are broadly predictable, but local changes like clouds are not. That makes it a legitimate and interesting project rather than a waste of effort, especially if you focus on improving adaptation.

A strong thesis angle would be combining camera control with image analysis, metering, or even short-term weather data to make exposure changes smarter and smoother.

Another possible approach is to keep a short fixed exposure and take multiple frames per interval in low light, then combine them into an equivalent longer exposure. That can work in some scenes, though noise and subject movement may limit it. Shooting RAW would help preserve dynamic range.

So: yes, systems like this do exist in principle, and building or improving one would be a sensible thesis topic.