A comparison of complex curve fitting and subspace identification algorithm for circuit modeling using frequency domain data

Efficient transient simulation of complex structures or components characterized by frequency-domain network parameters is of great interest in microwave theory. In this paper, two different model extraction techniques for broadband linear network characterization are discussed. A parametric complex curve fitting technique and a frequency-domain subspace identification algorithm are applied to generate passive circuit models for microwave devices characterized by measured frequency-domain port response data. By means of practical examples both methods are evaluated regarding usability, accuracy and CPU time.