A Note on \Book Review 'Tracking and Data Fusion: A Handbook of Algorithms'"

Y. Bar-Shalom, P. Willett and F. DaumFollowing the points raised in the review [1] of [2], an enlightening discussion took place, one that we feltwould be instructive to share with interested readers.The rst point brought up in [1] is the following regularity condition for the validity of the Cramer-Raolower bound (CRLB) for unbiased estimators of an unknown constant (nonrandom, or real valued) parameter,missing in [2]: \the support of the likelihood function [LF | the pdf of the observation zconditioned onthe parameter x, p(zjx)] must not depend on the parameter x. This technical condition is interestingand important, and perhaps its most familiar exemplar is that of the observations model of fz