Robot Controller

I am creating an FSM (Finite State Machine) that can be used as a controller to control the actions of a robot.

There is a grid with a food trail in it. My robot will have to find and follow that trail. There are gaps in the trail. The robot cannot keep a count of the food it found. It doesn't have to perfectly follow the entire food trace, and it shouldn't have masses of states to precise follow this food trail.

My available commands:

• move (1 step forward)
• left  (turns robot left)
• right  (turns robot right)
• food-ahead
• eat
• off-grid (determines if the robot has stepped off the grid)

I haven't decided yet if the robot will be able to return back to the grid in case that it has stepped off it.

Busy busy busy...

The first semester started and the modules are very interesting so far! =D I am taking an extra module C++ which is prerequisite for Games Technology for next semester ^^

I am currently looking for a data set so I can apply a data mining technique (such as clustering, association analysis, classification) to it. I would be interested in using Neural Networks. If so, I will have to separate my data set into two parts: the training and test set. After that I will train my network applying back propagation using the training set and check if everything is OK by running the test set on it. I should definitely be careful not to overfit it. That would be an interesting coursework. I have to investigate more my options though =D