Hello, I have aspirations to build a drone, but more pressingly I have an EE class in which I have to do some project at my discretion.

I decided to work in sensor fusion & localization, specifically to use a Kalman filter to get better readings about a ground based robot (I have the chassis built out of LEGO NXT parts, which use i2c and should be interoperable after a bit of magic).

After some reading, I feel comfortable with implementing a Kalman filter (specifically due to a good YouTube video by thescienceguy3000). I have not found a good 'first course' resource for complimentary filters and would appreciate being pointed to one.

Is the APM2.0 beefy enough to handle what I want it to do? How much of a delay should I expect after I 'preorder'?

PS: Its my understanding that a Kalman filtered magnetometer is not a good idea on a ground rover, since the assumption is of Gaussian noise, and in practice the noise is heavily biased off of the environment. Is there a practical way to make use of this sensor on a ground car?

PPS: What do I need to know that I neglected to ask about?