LEAST-SQUARES METHODS FOR H-INFINITY CONTROL-ORIENTED SYSTEM-IDENTIFICATION

This note presents a series of system identification algorithms that y ield identified models which are compatible with current robust contro ller design methodologies. These algorithms are applicable to a broad class of stable, distributed, linear, shift-invariant plants. The a pr iori information necessary for their application consists of a lower b ound on the relative stability of the unknown plant, an upper bound on a certain gain associated with the unknown plant, and an upper bound on the noise level. The a posteriori data information consists of a fi nite number of corrupted point frequency response estimates of the unk nown plant. The specific contributions of this note are to examine the extent to which certain standard Hilbert space or least squares metho ds are applicable to the H(infinity) system identification problem con sidered. Results are established that connect the H-2 error of the lea st squares methods to the H(infinity) error needed for control-oriente d system identification.