INFORMATION TYPE FILTERS FOR TRACKING A MANEUVERING TARGET

The structure of the information type Kalman filter can at times be used to develop numerically stable yet computationally efficient adaptive Kalman filters. Two previously proposed adaptive covariance type Kalman filtering techniques for tracking maneuvering targets are further developed to utilize the information type Kalman filter. These adaptive information type filters are described in structurally decoupled forms, thereby greatly reducing the computational requirements while rendering the filters amenable to implementation on parallel processors. A coherent decision procedure for selectively inducing partial coupling when necessary is developed via “off line” analysis of crosscorrelation functionals. © 1990 IEEE