Multistatic Tracking With the Precision Kalman Filter

This article presents a generalized, converted measurement Kalman filter for problems with a linear state equation and nonlinear measurement equation. This filter, called the precision Kalman filter (PKF), maps the measurements and their covariance matrices from measurement coordinates to state coordinates such that the converted measurements are unbiased and their covariance matrices are independent of the measurements. The filter’s performance is demonstrated on the multistatic sonar/radar tracking problems and is compared with that of the cubature Kalman filter (CKF) and another converted-measurement filter from the recent literature (called the CM2KF). These demonstrations show that the PKF has better overall performance than the CKF and CM2KF, particularly with respect to consistency.