Andrey Baev

Network Oriented Modeling of Passive Microwave Structures

Network oriented modeling of the passive microwave circuits provides a useful representation of the circuit and leads to the reduction of the required computer time and memory capacity. The proposed stability criterion allows dividing the simulation or measurement results into two parts representing a lumped elements model and distributed parameters model. The system identification procedure defines the order of the lumped elements model and the positions of poles and zeros in the complex frequency plane. The time-domain simulation results for the input impedance of the bowtie antenna in comparison with a method of moments solution are presented.

Near-field propagation of cyclostationary stochastic electromagnetic fields

We have extended the method for modeling the stochastic EM near-field which has already been described in [5, 13] for stationary stochastic fields to the case of cyclostationary fields. Areas of application are the modeling of the electromagnetic interference radiated by digital circuitry inside the system and also into the environment, where the period of the cyclostationary EMI is given by the clock frequency of the digital circuits.

Planar stochastic sources localization algorithm in EMC problems

A two dimensional localization algorithm for stochastic EMI sources based on time domain measurements of electromagnetic near-field tangential components is presented. The autocorrelation and cross-correlation functions obtained by using the two-point planar scanning system are used for the characterization of stochastic field distribution. The simulation and experimental data are processed for a spatial localization of equivalent point sources. The estimated field distributions and patterns are presented.

Transfer function representation of passive electromagnetic structures

The network-oriented ultra wideband transfer func- tion representation of complex three-dimensional electromagnetic structures is investigated. The transfer function consists of two parts: entire function and rational or pole function. System identification is carried out by the Matrix Pencil Method (MPM) and a new criterion for the choice of the model order and the late- time border of the impulse response is presented. This criterion is based on a pole set comparison by using the E-pulse method. The presented method yields considerable reduction of computational effort and allows to generate compact models of electromagnetic systems. I. Introduction The necessity of novel and broadband circuitry in nowadays applications, especially in mobile communication and vehic- ular sensing equipments, demands the availability of reliable and accurate modeling of complex electromagnetic structures up to the millimeter wave range. The simulation of these structures results in high computational effort and thus large simulation time. Therefore the application of network-oriented modeling in connection with system identification techniques allows reducing the computational effort and time by orders of magnitude. The transmission line matrix (TLM) method (1) and the finite-difference time-domain (FDTD) method provide a full- wave analysis of complex three-dimensional electromagnetic structures made of arbitrary materials. Their advantage is to characterize broadbandly a structure of N ports in only N simulation runs. This makes them the method of choice for e.g. ultra wideband devices. According to the singularity expansion method (3), the signal, scattered by the electromagnetic structure, can be decomposed into early-time and late-time parts. The early-time reaction is limited to a finite time interval and consists of the forced oscillations. This part can be represented by the entire function, which depends on the driving signal. The late-time response contains only natural oscillations but the late-time border can not be clearly defined by the structures parameters. The new method for late-time border determination is based on the phenomenon that the pole positions do not change for different slices of the late-time part of the impulse response (LTIR) if this part is fully described by the pole model. The deconvolution in time-domain can be done by matrix in- version, but the matrix obtained in this way is ill-conditioned. The rank reduction of an ill-conditioned matrix is performed in accordance with the spectral criterion (4). This criterion needs to take into consideration only the spectral components of the driving signal, exceeding some threshold within the frequency bandwidth of the electromagnetic structure. The application of the Matrix Pencil Method (MPM) for estimating the transfer functions of complex three-dimensional structures requires a strategy to choose the model order. The number of poles for the MPM estimation can be found by comparing the ratio of every singular value to the largest one with a predefined threshold (5). Such a choice can lead to an excessive number of poles. In this case some pole positions will change when different LTIR slices are used. A stable pole set can be defined by the stability criterion that is based on the comparison of pole sets via the E-pulse technique (6). This paper is organized as follows: First, the theory of transfer function representation is discussed, followed by the time-domain approach to the deconvolution problem for the impulse response reconstruction. After that the pole model of the system description and criteria for choosing the model order are considered. Further, the E-pulse comparison tech- nique is introduced. Finally, the MPM and E-pulse technique are applied to the time domain simulation data of a coplanar resonator structure operating at 80 GHz that is designed for material characterisation.

Application of the Stability Criterion to the Passive Electromagnetic Structures Modeling

The full wave time-domain numerical simulation of the 3-D electromagnetic structure yields wide band reaction of the multiport microwave circuit on the input electromagnetic signal. The realization of the system identification procedure demands the preliminary determination of the late-time border and the order of the pole model. The new criterion for their estimation is offered. This criterion is based on the stability of physical poles positions and uses the signature comparison technique as a tool of the distinction estimation. The impulse response reconstruction approach using the late time part of scattered waveform improves accuracy for parameter estimation in comparison with singular value decomposition (SVD) conventional procedure. The verification of the developed model for coplanar resonator is presented

TLM Modeling and System Identification of Distributed Microwave Circuits and Antennas

In this paper we present the time domain computation of the impulse response of optimized antenna structures using the transmission line matrix TLM method. Hybrid extensions based on system identification (SI) are a means to reduce the computational burden considerably. Model-based SI procedures can be used to represent complex electromagnetic structures by means of both lumped element networks and distributed parameters models.

Ultra wideband radar target discrimination using the signatures algorithm

An aspect independent radar target discrimination method based on the natural frequencies of ultra wideband radar targets is introduced. The sets of points in multi-dimension space corresponding to the poles on a complex plane (natural resonances) were offered as signatures of radar targets. This approach allows performing an automated discrimination algorithm of radar targets. The results of experimental research of the signals scattered by the scaled aircraft models by using the signatures algorithm are presented.

Parametric reconstruction of radar image based on Multi-point Scattering Model

The wideband coherent-pulse radar provides high-resolution image of the target. The model of this image is a complex envelope superposition corresponding to signals diffracted by the point scatterers. The values of complex envelopes are distributed over the radar image coordinate plane in accordance with the point scatterer positions and their reflection coefficients. The radar image model consists of range and Doppler profiles. The parameters of the target point scatterers were defined by processing of two-dimensional (2D) data extracted from the complex 2D discrete Fourier transforms of the radar image. The proposed parametric system identification method performs the estimation of the model parameters for a short dwell time and the extrapolation of the radar data image beyond this time. The modified procedure of inverse synthesis aperture radar imaging applied to actual data showed a reduction of the Doppler smearing and some improvements of image resolution.

Equivalent circuit model for coupled monolithic integrated millimeter-wave folded antennas

Network oriented modeling of passive microwave multiports provides a useful representation of the linear device and leads to the reduction of the required computer time and memory capacity for the synthesis procedure. The Brunes multiport method allows to represent the microwave structure in the predefined frequency band as a combination of the lossless connection ideal transformers network and lossy lumped elements with a minimum number of reactive elements. The system identification procedure defines the order of the lumped elements model and the positions of poles and zeros in the complex frequency plane. The simulation of the scattering parameters for the two-port formed by two coupled folded mm-wave antennas for automotive radar applications shows a good agreement with the corresponding measurements.

The joint inverse filtering and parametric identification for complex radar image

The paper suggests combining the inverse filtering and the parametric identification in the post-processing sequence of the complex radar image. The image can be modeled assuming the superposition of the identical partial responses from the effective target point-scatterers. Their positions correspond to the geometrical profile of target in the image coordinate plane. The proposed algorithm of radar image post-processing consists of the parametric and the non-parametric procedures which are used to improve the image resolution and to identify geometrical form of the target. The presented simulated results illustrate the main steps of radar image post-processing.