Optimal induced-norm and set membership state smoothing and filtering for linear systems with bounded disturbances

In this paper a unified framework founded on Information-Based Complexity i s introduced, to study set membership and optimal induced-norm state estima tion problems, for linear systems subject to norm bounded process noise and measurement errors. The proposed approach leads to a clean geometric pictu re of the problem, allowing for a straightforward derivation of several exi sting results. Moreover, it permits to tackle new estimation problems in wh ich both induced-norm optimization and consistency of the estimate with the noise bound are required. (C) 1999 Elsevier Science Ltd. All rights reserv ed.