Reduced order modelling of linear multivariable systems using an error minimization technique (Q1101388)

The method of reducing the order of a linear multivariable system is discussed. The dominant poles of the original system are retained, followed by matching the steady state parts of the unit step responses of the original and reduced systems. Each element of the transfer function matrix of the original system is considered separately. The coefficients of the numerator polynomials of the elements of the reduced order system transfer function matrix are then determined by minimizing the Integral Square Error (ISE) between the transient parts of the unit step reponses. An example illustrates the method and the result is compared to another method.