Generalised tracking
Nickens N. Okello, Daniel W. McMichael · 1999
Significant improvement in clutter and noise resistance can be gained from tracking both the kinematic and nonkinematic states of targets within a unified framework. The generalised tracking model performs centralised multisensor data fusion by propagating the posterior distribution of a generalised state vector comprising both continuous and discrete states. In applications, it can be used to unify tracking and identification. It performs better than approaches that track and identify separately, because it provides improved track-data association. The continuous variable segments of the generalised state vector can be tracked using any track-orientated algorithm, and the discrete segments are tracked using a hidden Markov model filter. This paper applies generalised tracking to the airborne early warning and control (AEW&C) surveillance problem to illustrate the integration of multiple dissimilar sensors.