On the identification of stochastic biases in linear time invariant systems

This paper addresses the existence of bias estimators. An approach to bias estimation is to augment the system state with bias states and implement a Kalman filter. Computational advantage can be gained using two parallel, reduced order Kalman filters. Conditions for existence of bias estimators for a linear, time invariant system with unknown, constant state and measurement biases are derived. A reduced row observability test matrix is used to show a necessary and sufficient condition for which complete bias observability does not exist. Examples are presented