A subspace approach for approximation of rational matrix functions to sampled data

Algorithms for approximation of rational matrix factors to data is described. The method is based on a subspace based multivariable frequency domain state-space identification, canonical and spectral factorization and parametric optimization. The algorithms can be used for identifying spectral factors and factors of positive real functions from frequency data. The methods are directly applicable in the D-K algorithm for complex /spl mu/-synthesis and the Y-Z-K algorithm for mixed /spl mu/-synthesis.