Characteristic impedance of surface-strip coplanar waveguides

Adaptive Neuro-Fuzzy Models for Conventional Coplanar Waveguides

“…There are small differences between the ANFIS model results and the experimental results. These differences are expected because it is assumed that the ground planes extend to infinity and the conductor thicknesses are zero for the ANFIS models, however, the CPW structures of the experimental works [24][25][26] have finite size ground planes in the range of 1270.0 µm and 3505.2 µm, and the CMT [3] ANFIS Measured [26] MMICTL [23] (c) S = 508 µm, h = 635 µm, and ε r = 9.6 Figure 5. Comparison of the measured and calculated characteristic impedances obtained by using ANFIS, CMT [3], and MMICTL [23] for CCPWs.…”

“…In order to further validate the presented method, the characteristic impedance test results of ANFIS model are compared with the MMICTL [23] and measured [24][25][26] results in Figure 5. It is clear from this figure that the ANFIS results agree quite well with the results of measured works and also with the results of MMICTL.…”

“…The adaptation is carried out after the presentation of each set (ε r , d/h and S/d) until the calculation accuracy of the ANFIS is deemed satisfactory according to some criterion (for example, when the error between target and the actual output for all the training set falls below a given threshold) or when the maximum allowable number of epochs is reached. The training and test data sets used in this article have been obtained from previous numerical methods [3,14], simulation [23] and experimental works [24][25][26] results. The number of epochs used for training was 150.…”

Very accurate and simple CAD models based on neural networks for coplanar waveguide synthesis

“…In these comparisons, the results obtained from the first and second neural models are compared with three other approaches from the literature. These are the results of the respective quasi-static analyses, two CAD models which are known as the best accurate tools for CPW synthesis available in the literature [14,17], and experimental work reported in [36].…”

“…These neural models were trained using the LM [29,30], Bayesian regularisation (BR) [31], conjugate gradient of Fletcher-Powel (CGF) [32][33], backpropagation with momentum (BPM) [34], and quasi-Newton (QN) [35] learning algorithms to obtain better performance and faster convergence with a simpler structure. For the validation of the neural models presented in this work, the synthesis results obtained from these neural models have been compared with the quasi-static analysis results [3,4], the results of the synthesis formulas proposed by other researchers [14,17], and the experimental results [36] available in the literature.…”