Abstract
An efficient minimal realization technique that is based on the dyadic expansion of transfer-function matrices is introduced. The proposed algorithm is fundamental and straightforward in the sense that it does not require the solution of sets of equations or any polynomial manipulations and it can therefore overcome the need for using the computer, even in problems of relatively high order. The cases of both distinct and multiple poles are rigorously treated and a detailed algorithm is given which is illustrated by two examples. A reduction in digital computation time of the order of 1 : 10 compared with other realization algorithms has been achieved.

This publication has 5 references indexed in Scilit: