Sunday, March 14, 2010

Newton vs Kalman; Estimating Motion

Surprisingly often I'm faced with estimating the position of something that moves according to Newtonian laws of motion in one dimension and with random acceleration.

A fast and pretty accurate way of doing this is to use a Kalman filter. I've finally gotten around to implementing a specialized version of this general method, dedicated to my simple estimations needs. You are welcome to use it (it's HPL licensed).

Have a look at the image if you want to see how it performs in a sample case.

No comments:

Post a Comment