Optimal Recursive Estimation with Uncertain Observation and Time-Correlated Disturbances

Mao Sheng
2005-01-01
Abstract:Considers the least mean-squared error linear estimation problem, using covariance information, in linear discrete-time stochastic systems with uncertain observations for the case where the state and measurement noises are correlated both at the same time and the neighbored time, while the state noises are not white. The different kinds of estimation problem treated include one-stage prediction and filtering. Employing the Orthogonal Projection Lemma and the features of the observations derives the recursive algorithm out. Simulation result shows the effectiveness of the algorithm.
What problem does this paper attempt to address?