> As far as i understand, Kalman Filters seem to be a moving average of position, velocity, and (perhaps) acceleration, and use those 3 values to estimate the 'real value' as opposed to what the sensor is telling you.

Sort of. The actual variables (v, r, a in your case) vary depending on your use case. What matters is you make an estimate based on that state, then you take a measurement and there’s an iterative way of adjusting your next state based on those two. It’s actually pretty easy to code once you understand formulas.

Maybe this could help http://bilgin.esme.org/BitsAndBytes/KalmanFilterforDummies