Linear algorithms for worst case identification in H∞ with applications to flexible structures

This paper is concerned with linear algorithms for identification in H^∞ proposed by Helmicki et al. (1993). It is shown that these two different linear algorithms can be unified into a single one which can be further extended to nonuniformly spaced frequency response samples with exponential convergence for the noise free case. Improved upper bounds for the corresponding identification errors are derived. Applications to the identification of lightly damped systems such as flexible structures are also considered.