Status of CR (Cramer-Rao)-Like Lower Bounds for Nonlinear Filtering.

摘要

The motivation and mechanics of utilizing Cramer-Rao (CR)-type lower bounds are reviewed as used to gauge the performance of filters being evaluated in nonlinear estimation applications such as in sonar/sonobuoy/radar target tracking. The status of several similar alternative CR-type lower bounds that have been considered or used for this purpose are reviewed and certain limitations/caveats associated with their use are offered. These results should be of interest to sonar/sonobuoy/radar practitioners and Kalman filter or nonlinear filter theorists. Reprints. (rrh)