I'm Mate Kisantal, an Aerospace Engineering MSc student at TU Delft, in the Netherlands. My master track is Control&Simulations, and my particular interest is in machine learning approaches for intelligent control.
In particular, I'm working on a Deep RL approach to enable drones to intelligently navigate and avoid obstacles in cluttered indoor/outdoor environments. This would allow greater autonomy and safer operation for drones, paving the way for autonomous drone delivery and a number of other interesting applications.
Since this sort of training is better done in a simulator than on real drones, one of my main research direction is domain adaptation. I'm trying to ensure that the skills and knowledge learnt with synthetic inputs can be transferred to the real world, and operate accurately on-board a real drone. But I see many opportunities to use simulators and synthetic data for training once effective domain adaptation is available, not only for drones but in general!
I'm interested to learn more about AI and ML, ant to see what research opportunities you guys see in this rapidly developing field!