Multivariabe adaptive control without a prior knowledge of the delay matrix (Q1091998)

This paper is concerned with the structure and role of the time-delay matrix in discrete-time multivariable processes. Interpretations and properties of this matrix as a multivariable generalization of the SISO delay term are discussed. It is formally shown that a finite-time horizon or multi-step cost function for the multivariable case can be minimized subject to a suitable choice of output and control horizons and without prior knowledge of the system delay matrix.