For example, in the transportation balloon competition held by the Ministry of Education two years ago, the positioning method and target recognition method of intelligent robots must be original in theory. From the video, it seems that the movements are similar. In fact, by consulting their published literature, we can find that almost all different teams have put forward their own theories. For identifying the circle of scoring points, some use video edge detection theory, and some collect images to establish functional Bayesian estimation theory. Only amateurs can learn and watch this kind of competition, because it was originally made by masters from all fields of participating colleges with their students. Even if you find a doctoral supervisor in a certain direction to do it, you can only complete that part of his field.
If you are interested in this, I suggest that you might as well participate in some groups and competitions of folk robot lovers and have some exchanges. You can join the assembly robot enthusiast groups produced by some manufacturers, or you can join some robot enthusiast teams for non-professionals. I don't know much about these situations. Last year, the education department of Harbin established the Robot Guidance and Competition Association for primary and secondary schools. For example, Lego mindstorms has some groups in China (this is not an advertisement, there are many other brands). I suggest you talk to a similar local association.