MaxEnt useable to find `best fitting' Gaussian?

Herman Bruyninckx (Herman.Bruyninckx@mech.kuleuven.ac.be)
Mon, 1 Feb 1999 14:34:32 +0100 (CET)

I have a system (RoboCup, i.e., a simulation of soccer playing robots) that
uses uniform distributions to model the ``noise'' on the motion of the
players and the ball. I want to apply a Kalman FIlter to track these moving
objects, so I need the ``best'' way to approximate the uniform
distribution by a Gaussian distribution? Should (can) I use the MaxEnt
principle to find this? If yes, what are the ``constraints'' imposed by the
uniform distribution?

--
Herman.Bruyninckx@mech.kuleuven.ac.be (Ph.D.)    Fax: +32-(0)16-32 29 87
Dept. Mechanical Eng., Div. PMA, Katholieke Universiteit Leuven, Belgium
        <http://www.mech.kuleuven.ac.be/~bruyninc>