Descriptor Systems Kalman Fusion Predictor based on Covariance Intersection

For descriptor systems with unknown covariance, a covariance intersection (CI) fusion robust steady-state Kalman predictor is presented, by using the descriptor value decomposition ，the problems of descriptor systems changed into the problems of two normal systems. By using covariance intersection (CI) fusion , a reduced order covariance intersection (CI) fusion steady-state descriptor Kalman predictor is proposed. A Monte-Carlo simulation example for a tracking system shows that its actual accuracy is inferior to that of the optimal Kalman fuser and is prior to local Kalman predictor of single sensor.