Parametric modeling of microwave filter using combined MLS-SVR and pole-residue-based transfer functions

摘要

In this article, an advanced technique is developed to combine multi-output leastsquaresrnsupport vector regression (MLS-SVR) and pole-residue-based transfer functionrnmodels for microwave filter parametric modeling. MLS-SVR is trained to learnrnthe relationship between the length of tuning screws and the pole/residues of therntransfer function, where MLS-SVR is an effective method to cope with the multioutputrncase unlike the traditional approach. Traditional approach treats the differentrnoutputs separately in the multi-output case and it cannot model the relation betweenrndifferent outputs. Another important element for modeling is feature parameters.rnExtracted feature parameters have an important influence on the accuracy of modeling.rnFor the purpose of establishing a more accurate model, the complex system polesrnand residues from Y-parameters are chosen as the outputs of modeling, which arernobtained by vector fitting (VF). Then we give a solution to obtaining pole/residuesrnextracted by VF when the filter is in high detuned. After the proposed modeling process,rntrained model can be used to provide an accurate and fast prediction of thernbehavior of microwave filter with the length of tuning screws as variables, and modelrnthe electromagnetic simulation (or actual) microwave filter tuning. The methodologyrnis applied to a narrow band coaxial-resonator filter modeling, and more accuraternresults are achieved compared with the other methods. An example is presented tornillustrate the efficiency of the proposed method.