Time filter

Source Type

Michalski J.J.,TeleMobile Electronics Ltd.
4th Microwave and Radar Week, MRW-2010 - 18th International Conference on Microwaves, Radar and Wireless Communications, MIKON 2010 - Conference Proceedings | Year: 2010

This paper shows how to improve the efficiency of the Artificial Neural Network (ANN) method for the cavity filter tuning. It was proved that the usage of many golden filters in the process of collecting the learning vectors, used in ANN training, has the significant influence in decreasing the ANN generalization error. Thus, the algorithm efficiency is increasing. The generalization error value of ANN, trained on samples from two different filters, as a norm of the filters similarity is proposed. The tuning experiment for the 6-cavities RX filters of GSM diplexer has been demonstrated. In the experiment the ANNs were trained based on the vectors collected from up to five different filters, showing the significant influence of the number of "known filters" on the ANN generalization error.

Michalski J.J.,TeleMobile Electronics Ltd. | Kowalczyk P.,Technical University of Gdansk
IEEE Transactions on Microwave Theory and Techniques | Year: 2011

This paper presents a novel method that is very efficient in solving multidimensional real and complex eigenvalue problems, commonly employed in electromagnetic analysis, which can be transformed into a nonlinear equation. The concept is realized as root tracing process of a real or complex function of N variables in the constrained space. We assume that the roots of the continuous function of N variables lie on the continuous (N-1)-dimensional hyperplane. The method uses regular N and (N-1)-Simplexes, at which vertices the considered function changes its sign. Based on (N-1)-Simplex, the function is evaluated at two new points that are vertices of new regular N-Simplexes for which (N-1)-Simplex is one of its (N-1)-faces. The algorithm, with the usage of stack, runs in an iterative mode tracing the roots inside the volume of the considered simplexes. As a result, the algorithm creates a chain of simplexes in the constrained region. The proposed algorithm is optimal in the sense of the number of function evaluations. The numerical results, real and complex dispersion characteristics of chosen microwave guides, have proven the versatility and efficiency of the proposed algorithm. © 2011 IEEE.

Michalski J.J.,TeleMobile Electronics Ltd.
Progress in Electromagnetics Research | Year: 2011

This paper presents a new method of sequential microwave filter tuning. For filters with R tuning elements (including cavities, couplings and cross-couplings), based on physically measured scattering characteristics in the frequency domain, the Artificial Neural Network (ANN) is used to build inverse models of R sub-filters. Each sub-filter is associated to one tuning element. The sub-filters are obtained by successive opening or shorting of resonators and by removing coupling screws. For each sub-filter, the ANN training vectors are defined as physical reflection characteristics (input vectors) and the corresponding positions of the tuning element, which is detuned, in both directions, from its proper setting (output vectors). In the tuning process, such inverse models are used for calculating the tuning element increments needed for setting the tuning element in the proper position. The tuning experiment, conducted on 8- and 11-cavity filters, has shown the performance of the presented method.

Michalski J.J.,TeleMobile Electronics Ltd.
Progress In Electromagnetics Research M | Year: 2010

This paper presents a novel method of cavity filter tuning with the usage of an artificial neural network (ANN). The proposed method does not require information on the filter topology, and the filter is treated as a black box. In order to illustrate the concept, a feed-forward, multi-layer, non-linear artificial neural network with back propagation is applied. The method for preparing, learning and testing vectors consisting of sampled detuned scattering characteristics and corresponding tuning screw deviations is proposed. To collect the training vectors, the machine, an intelligent automatic filter tuning tool integrated with a vector network analyzer, has been built. The ANN was trained on the basis of samples obtained from a properly tuned filter. It has been proved that the usage of multidimensional approximation ability of an ANN makes it possible to map the characteristic of a detuned filter reflection in individual screw errors. Finally, after the ANN learning process, the tuning experiment on 6 and 11-cavity filters has been preformed, proving a very high efficiency of the presented method.

Kacmajor T.,TeleMobile Electronics Ltd. | Michalski J.J.,TeleMobile Electronics Ltd.
IEEE MTT-S International Microwave Symposium Digest | Year: 2011

This paper describes a method of microwave filter tuning. The main goal of this research is to take advantage of fuzzy logic and build effective, adaptive approximator - a neuro-fuzzy system. The system was trained with use of samples that contain information about scattering characteristics and corresponding tuning screw deviations. Experiments were performed on four different filters which have 6 to 11 cavities. The results have been then compared with previous works which use artificial neural networks. The system learning phase has been proved to reach lower generalization and learning error. © 2011 IEEE.

Discover hidden collaborations