In my machine learning course, our professor wants us to kind of lead our own projects this semester in contrast to having given very specific assigned projects in the past. It is our task to come up with a cool machine learning problem and then go out and actually solve it.
So, over the past few days I have had a lot of things run through my mind...a lot of different ideas. One of the ideas is Lego Mindstorms. I think it would be really cool to construct a Lego Mindstorms robot and do some machine learning with it, but for the life of me I can't think of a good problem to solve using a Lego Mindstorms robot.
So I was looking for ideas, and I thought maybe you guys could help me out. What kind of cool machine learning problems could you solve with a Mindstorms robot? I'm not talking about the classic "follow a dark line" or "get all the cans out of the circle". Those can easily be done without any "learning" at all.
I need a problem that requires the task of actually gathering lots of data, and then performing operations of classification, clustering, etc. Something that use perceptrons, decision trees, support vector machines, etc. etc. etc.