A canonical transform is presented that converts a coupled or uncouple
d kinematic model for target tracking into a decoupled dimensionless c
anonical form. The coupling is due to non-zero off-diagonal terms in t
he covariance matrices of the process noise and/or the measurement noi
se, which can be used to model the coupling of motion and/or measureme
nt between coordinates. The decoupled dimensionless canonical form is
obtained by simultaneously diagonalizing the noise covariance matrices
, followed by a spatial-temporal normalization procedure. This canonic
al form is independent of the physical specifications of an actual sys
tem. Each subsystem corresponding to a canonical coordinate is charact
erized by its process noise standard deviation, called the maneuver in
dex as a generalization of the tracking index for target tracking, whi
ch characterizes completely the performance of a steady-state Kalman f
ilter. A number of applications of this canonical form are discussed.
The usefulness of the canonical transform is illustrated via an exampl
e of performance analysis of maneuvering target tracking in an air tra
ffic control (ATC) system.