SEPARATE-BIAS KALMAN ESTIMATOR WITH BIAS STATE NOISE

The problem of estimating a set of dynamic states and a set of slowly varying bias-like states employing a decoupled Kalman estimator approach is addressed. A decoupled estimator structure, suitably modified from that developed by Friedland in dealing with the constant-bias case, is shown to have the potential of essentially optimal performance when the bias vector undergoes limited random variation. © 1990 IEEE