Convolutional Neural Network for Coupling Matrix Extraction of Microwave Filters (original) (raw)

Classification of Microwave Planar Filters by Deep Learning

Radioengineering, 2022

Over the last few decades, deep learning has been considered to be powerful tool in the classification tasks, and has become popular in many applications due to its capability of processing huge amount of data. This paper presents approaches for image recognition. We have applied convolutional neural networks on microwave planar filters. The first task was filter topology classification, the second task was filter order estimation. For the task a dataset was generated. As presented in the results, the created and trained neural networks are very capable of solving the selected tasks.

Design of a Microwave Lowpass – Bandpass Filter using Deep Learning and Artificial Intelligence

Journal of the Institute of Electronics and Computer, 2021

In this paper, a lowpass – bandpass dual band microwave filter is designed by using deep learning and artificial intelligence. The designed filter has compact size and desirable pass bands. In the proposed filter, the resonators with Z-shaped and T-shaped lines are used to design the low pass channel, while coupling lines, stepped impedance resonators and open ended stubs are utilized to design the bandpass channel. Artificial neural network (ANN) and deep learning (DL) technique has been utilized to extract the proposed filter transfer function, so the values of the transmission zeros can be located in the desired frequency. This technique can also be used for the other electrical devices. The lowpass channel cut off frequency is 1 GHz, with better than 0.2 dB insertion loss. Also, the bandpass channel main frequency is designed at 2.4 GHz with 0.5 dB insertion loss in the passband.

A new neural network technique for the design of multilayered microwave shielded bandpass filters

International Journal of Rf and Microwave Computer-aided Engineering, 2009

In this work, we propose a novel technique based on neural networks, for the design of microwave filters in shielded printed technology. The technique uses radial basis function neural networks to represent the non linear relations between the quality factors and coupling coefficients, with the geometrical dimensions of the resonators. The radial basis function neural networks are employed for the first time in the design task of shielded printed filters, and permit a fast and precise operation with only a limited set of training data. Thanks to a new cascade configuration, a set of two neural networks provide the dimensions of the complete filter in a fast and accurate way. To improve the calculation of the geometrical dimensions, the neural networks can take as inputs both electrical parameters and physical dimensions computed by other neural networks. The neural network technique is combined with gradient based optimization methods to further improve the response of the filters. Results are presented to demonstrate the usefulness of the proposed technique for the design of practical microwave printed coupled line and hairpin filters. © 2008 Wiley Periodicals, Inc. Int J RF and Microwave CAE, 2009.

Effective Design of Cross-Coupled Filter Using Neural Networks and Coupling Matrix

2006 IEEE MTT-S International Microwave Symposium Digest, 2006

In this paper, neural network modeling techniques are applied to the design of waveguide dual-mode pseudo-elliptic filter. A hybrid modeling approach is developed where neural networks and filter coupling matrix are combined in an innovative way to deliver speed and accuracy of the overall filter design. Filter structure is decomposed into modules representing each coupling mechanism. Generalized scattering matrices (GSM) of the modules are calculated using mode-matching method. Equivalent circuit parameters, such as coupling value and insertion phase lengths are then extracted from EM data. Neural models are developed for circuit parameters for each individual module instead of direct modeling of GSM. Good agreement is obtained between neural models and EM-based data. A narrow-bandwidth four-pole Ku band bandpass filter is designed using the trained NN models, and simulated and optimized using full EM model (HFSS). The difference between the optimized dimensions and NN model is within 0.01" for all dimensions, which demonstrates that the developed NN models are capable of achieving the accuracy of EM-based model with superior computation speed. Index Terms-Bandpass filters, dual mode filters, coupling matrix, microwave filters, neural networks.

High-Dimensional Neural-Network Technique and Applications to Microwave Filter Modeling

IEEE Transactions on Microwave Theory and Techniques, 2010

Neural networks are useful for developing fast and accurate parametric model of electromagnetic (EM) structures. However, existing neural-network techniques are not suitable for developing models that have many input variables because data generation and model training become too expensive. In this paper, we propose an efficient neural-network method for EM behavior modeling of microwave filters that have many input variables. The decomposition approach is used to simplify the overall high-dimensional neural-network modeling problem into a set of low-dimensional sub-neural-network problems. By incorporating the knowledge of filter decomposition with neural-network decomposition, we formulate a set of neural-network submodels to learn filter subproblems. A new method to combine the submodels with a filter empirical/equivalent model is developed. An additional neural-network mapping model is formulated with the neural-network submodels and empirical/equivalent model to produce the final overall filter model. An -plane waveguide filter model and a side-coupled circular waveguide dual-mode filter model are developed using the proposed method. The result shows that with a limited amount of data, the proposed method can produce a much more accurate high-dimensional model compared to the conventional neural-network method and the resulting model is much faster than an EM model.

Artificial Neural Network Approach for Modeling the Microstrip Bandpass Filter

Artificial neural networks recently gained attention as a fast and flexible vehicle to microwave modeling and design. The advantage of using ANN models for filter design is a large saving in required CPU time. In this paper, an end coupled microstrip band-pass filter is designed to operate in the frequency range of 4GHz to 6GHz and its S-parameter responses are developed using ANN. Here the filter characteristics are analyzed using different learning algorithms like back propagation (MLP3), Quasi-Newton (MLP), Sparse training, Huber-Quasi-Newton, adaptive back propagation, auto pilot (MLP3) and conjugate gradient, and the results are compared with the simulated results in terms of correlation coefficient and average error.

Automated Modeling of Microwave Structures by Enhanced Neural Networks

2006

The paper describes the methodology of the automated creation of neural models of microwave structures. During the creation process, artificial neural networks are trained using the combination of the particle swarm optimization and the quasi-Newton method to avoid critical training problems of the conventional neural nets. In the paper, neural networks are used to approximate the behavior of a planar microwave filter (moment method, Zeland IE3D). In order to evaluate the efficiency of neural modeling, global optimizations are performed using numerical models and neural ones. Both approaches are compared from the viewpoint of CPU-time demands and the accuracy. Considering conclusions, methodological recommendations for including neural networks to the microwave design are formulated.

Prior knowledge based neural modeling of microstrip coupled resonator filters

Facta universitatis. Series electronics and energetics, 2022

The design of microstrip coupled resonator filters includes determination of the coupling coefficients between the filter resonator units. In this paper a novel modeling procedure exploiting prior knowledge neural approach is proposed as an efficient alternative to the standard electromagnetic (EM) simulations and to the neural models based purely on the artificial neural networks (ANNs). It has similar accuracy as the EM simulations and requires less training data and less time needed for the model development than the models based purely on ANNs.

Application of neural networks: Enhancing efficiency of microwave design

The paper describes the methodology of the automated creation of neural models of microwave structures. During the creation process, artificial neural networks are trained using the combination of the particle swarm optimization and the quasi-Newton method to avoid critical training problems of the conventional neural nets. Neural models are required being wideband. In order to meet this requirement, feed-forward neural networks and recurrent ones are used for modelling, and their properties are in detail mutually compared. In the paper, neural networks are used to approximate behaviour of a planar microwave filter (moment method, Zeland IE3D). In order to evaluate the efficiency of neural modelling, global optimizations are performed using numerical models and neural ones. Both approaches are compared from the viewpoint of CPU-time demands and accuracy. Considering conclusions, methodological recommendations for including neural networks to microwave design are formulated.

Neural Network Inverse Modeling and Applications to Microwave Filter Design

IEEE Transactions on Microwave Theory and Techniques, 2008

In this paper, systematic neural network modeling techniques are presented for microwave modeling and design using the concept of inverse modeling where the inputs to the inverse model are electrical parameters and outputs are geometrical parameters. Training the neural network inverse model directly may become difficult due to the nonuniqueness of the input-output relationship in the inverse model. We propose a new method to solve such a problem by detecting multivalued solutions in training data. The data containing multivalued solutions are divided into groups according to derivative information using a neural network forward model such that individual groups do not have the problem of multivalued solutions. Multiple inverse models are built based on divided data groups, and are then combined to form a complete model. A comprehensive modeling methodology is proposed, which includes direct inverse modeling, segmentation, derivative division, and model combining techniques. The methodology is applied to waveguide filter modeling and more accurate results are achieved compared to the direct neural network inverse modeling method. Full electromagnetic simulation and measurement results of -band circular waveguide dual-mode pseudoelliptic bandpass filters are presented to demonstrate the efficiency of the proposed neural network inverse modeling methodology.