International Workshop on

Advances in Modeling and Optimization
of High Frequency Structures

Confirmed speakers (in alphabetical order):

• Mohamed Bakr (McMaster University) “Time Domain Sensitivity Analysis and Their Applications: State of the Art” Abstract

• John W. Bandler (Bandler Corporation, McMaster University) “Space Mapping: Physics-Driven Optimization Technology for Effective Engineering Modeling and Design” (co-authors: S. Koziel, Q.S. Cheng, and K. Madsen) Abstract

• Stéphane Bíla (Université de Limoges) “An Overview on EM-Based Computer-Aided Design Techniques for Microwave Devices” Abstract

• Bruno Carpentieri (University of Groningen) “Recent Advances in Fast Integral Solution Methods for Electromagnetic Scattering Analysis of High Frequency structures” Abstract

• Santiago Cogollos (Universidad Politécnica de Valencia) “Recent Advances for the Full-Wave Analysis and Design of Waveguide filters for Space Applications” (co-authors: Pablo Soto, Vicente E. Boria, Carlos Vicente, Jordi Gil, and Benito Gimeno) Abstract

• Tom Dhaene (Ghent University) “Surrogate-based modeling of electrical systems” (co-author: I. Couckuyt) Abstract

• Hans Hjelmgren (Chalmers University of Technology) “Numerical simulation of microwave power transistors” Abstract

• Slawomir Koziel (Reykjavik University) “Response Correction Methods for Microwave Design Optimization” Abstract

• Giuseppe Macchiarella (Politecnico di Milano), “Advanced CAD techniques for Filters and Combiners in Wireless Base Stations” Abstract

• Fermín Mira (Centre Tecnològic de Telecomunicacions de Catalunya), “Wideband Representation of Passive Components Based on Planar Waveguide Junctions” (co-authors: Ángel A. San Blas, Vicente E. Boria, and Benito Gimeno) Abstract

• Natalia Nikolova (McMaster University) “Microwave Real-time Detection of Scatterers Using Self-adjoint Sensitivity Analysis” Abstract

• Stanislav Ogurtsov (Reykjavik University) “Design Optimization of Planar Antennas for UWB Communication” (co-author: S. Koziel) Abstract

• James Rautio (Sonnet Software Inc.) “Examples of Microwave Filter Optimization Using Perfectly Calibrated Ports” Abstract

• C.J. Reddy (EM Software & Systems (USA) Inc.) “Optimization of High Impedance Surfaces for Compact Planar Antenna Design” (co-authors: Gopinath Gampala and Rohit Sammeta) Abstract

• Vittorio Rizzoli (Universitá di Bologna) “Nonlinear/electromagnetic co-simulation of microwave/millimeter-wave radio systems” (co-authors: Alessandra Costanzo, Diego Masotti, and Francesco Donzelli) Abstract

• Peter Russer (Technische Universitaet Muenchen), “Network methods for electromagnetic field and multiphysics modeling” Abstract

• Rolf Schuhmann (University of Paderborn) “Analysis of parametrized eigenvalue problems for dielectric waveguides” (co-author: Andrea Walther) Abstract

• Almudena Suárez (University of Cantabria) “Recent advances in stability analysis of power amplifiers” (co-authors: Franco Ramirez, Sanggeun Jeon, and David Rutledge) Abstract

• Peter Thoma (CST AG), “Using Coupled Simulations to Efficiently Address Challenging EM Problems” Abstract

• Manos M. Tentzeris (Georgia University of Technology) “Design and optimization of liquid antennas for biomonitoring applications” (co-author: A. Traille) Abstract

• Manos M. Tentzeris (Georgia University of Technology) “Inkjet-Printed Paper/Polymer-Based "Green" RFID and Wireless Sensor Nodes: The Final Step to Bridge Cognitive Intelligence, Nanotechnology and RF” Abstract

• Qi-Jun Zhang (Carleton University), “Neural Networks for High-Frequency Component Modeling” Abstract

    

---

       Main Page

---

---

Abstract: We discuss in this presentation the recent advances in the area of time-domain sensitivity adjoint analysis of high frequency structures.  These techniques aim at efficiently estimating the sensitivities of a desired response with respect to all parameters of interest using a time domain simulator.  Classical approaches obtain these sensitivities using finite difference approaches.  For a structure with n parameters, at least n extra simulations are required.  Adjoint sensitivity analysis approaches obtain the required sensitivities using at most one extra simulation.  In some cases, no extra simulations are required and the problem is denoted as a self-adjoint problem.   We focus on recent approaches that allow for estimating the sensitivities of the complete transient response at all time steps with respect to all parameters.  The transient sensitivities are estimated using at most on extra simulation regardless of the number of parameters and regardless of the number of time steps.  We illustrate the application of these approaches to the solution of inverse problems.

List of talks    Main Page

---

Abstract: Engineering design optimization through space mapping addresses the iterative enhancement of a suitable physics-based surrogate of the target structure.  Such surrogates include fast-to-compute, physically-based “coarse” models that represent expensive “fine” or “high-fidelity” models relatively well.  In such cases, a space mapping algorithm provides excellent designs after only a few high-fidelity simulations.
The methodology follows the engineer’s traditional intuition, yet is amenable to mathematical treatment.  It mimics the way the brain relates new objects or images with familiar objects, images, reality, or experience and offers a quantitative explanation for an expert’s “feel” for a physical problem.  It takes the high-fidelity simulator out of the classical optimization loop, relegating classical optimization algorithms to inexpensive coarse or surrogate models.  Starting from the early days of the Newton-like aggressive space mapping algorithm, the methodology has an excellent track record of success in diverse engineering areas: electronic, photonic, antenna, microwave and magnetic systems; civil, mechanical, and aerospace engineering structures.
Here, we consider efficient EM-based design suggested by recent results on design closure, port tuning, and perfectly calibrated ports.  We mention space-mapping-based modeling and optimization, both interacting concurrently with full-wave EM solvers, and we indicate recent advances in the so-called “tuning space mapping” process.  We suggest why, when, and how the methodology works.  Microwave engineering examples involving commercial EM solvers and some recent IEEE Microwave Magazine articles complement our presentation.

List of talks    Main Page

---

Abstract: The exponential growth of data exchanged by information and communication networks requires new equipments with improved performances and functionalities. Regarding the application, integration, durability and cost are particular constraints to be adjusted during the design.
Telecommunication systems benefited from technological advances performed in electronics and computer science. At the same time, sophisticated computer-aided design techniques have been developed and, with the support of powerful computers, became more and more popular in engineer’s offices. Nowadays, CAD software establish themselves as necessary tools for innovation in order to develop new components.
CAD techniques are often developed for a specific usage; but the problems and, as a result, the techniques used for modeling and optimizing microwave components often meet.
The presentation offers an overview of EM-based CAD techniques used for the design of microwave devices. The talk will be focused on the design of microwave filters and will be obviously non-exhaustive, but several approaches related to model order reduction and driven optimization techniques will be discussed.
Several examples will illustrate the presentation and a particular place will be devoted to shape optimization techniques.

List of talks    Main Page

---

Abstract: An efficient solution of the Maxwells equations is a critical component of the simulation of many realistic industrial processes e.g. the Radar Cross Section (RCS) calculation of arbitrarily shaped electrically large objects like aircrafts, the analysis of electromagnetic (EM) compatibility
of electrical devices with their environment, the design of absorbing materials, radars, antennas and many others. Objects of interest in real-life applications can have large electrical size compared to the wavelength. Therefore, the numerical solution is extremely demanding for large computer
resources and fast numerical algorithms.
Impressive advances in computer technology have made a rigorous numerical solution of high frequency scattering problems possible for many practical applications. Differential equation solvers are a popular solution approach for the Maxwells equations. However, alternative approaches based on integral equation methods are getting ground. Integral methods solve for the electric and the magnetic currents induced on the surface of the object and require a simple description of the surface of the target by means of triangular facets, so that a 3D problem is reduced to solving a
2D surface problem simplifying considerably the mesh generation especially in the case of moving objects.
In this presentation we overview trends and problems in the design of a practical surface integral equation solver for electromagnetic scattering analysis of high frequency structures. We focus our attention to the solution of the large and dense linear system arising from the discretization, that is one bottleneck of the computation. Scattering simulations may involve meshes with several million points and the memory requirements of out-of-core dense direct solvers are not aordable even on modern parallel computers.
We address algorithmic development, implementation details, numerical scalability, parallel performance. Thanks to the use of iterative methods and suitable preconditioners, fast integral solvers involving tens of million unknowns are nowadays feasible and can be integrated in the design processes.

List of talks    Main Page

---

Abstract: Some recent advances in waveguide filter analysis and design are presented throughout this talk. These advances have been focused in the filter classes that are more commonly used in the satellite hardware (e.g. dielectric filters, lowpass filters, bandpass filters, dual mode filters and tunable filters). Sections presented in this talk correspond to the fields in the EM-based filter design area that have been improved or where new contributions have been added: i.e. analysis, synthesis and optimization. These new methods are being incorporated in the commercial tool FEST3D (Full-wave Electromagnetic Simulation Tool for 3D waveguide components) giving better flexibility, increased accuracy and higher speed in the design process.

List of talks    Main Page

---

Abstract: As prototype building is very costly, the use of computer simulations has become commonplace as a feasible alternative. However, due to the computational cost of these high-fidelity simulations, the use of behavioral or surrogate modeling techniques have become indispensable.
Surrogate models are compact and cheap to evaluate, and have proven very useful for tasks such as optimization, design space exploration, prototyping, and sensitivity analysis. Consequently, there is great interest in techniques that facilitate the construction of such behavioral models, while minimizing the computational cost and maximizing model accuracy. We present an algorithm that integrates adaptive modeling and adaptive sampling methods in order to generate an accurate global approximation over the design space of interest with a minimum number of electromagnetic simulations. Moreover, the presented approach does not mandate assumptions (but does not preclude them either) about the problem, i.e., the electromagnetic simulator is viewed as a black- box.

List of talks    Main Page

---

Abstract: There is a large interest in high efficient power amplifiers for microwave and millimeter wave frequencies.  Power transistors made on silicon carbide (SiC) and gallium nitride (GaN) can deliver higher power per unit width than silicon and GaAs (gallium arsenide) transistors. Large signal steady-state characteristics are easily obtained from harmonic balance simulations of equivalent circuit transistor models in the frequency domain. This can be done in numerical simulation of semiconductor devices (TCAD) as well, but the simulation time becomes considerable. As an alternative large signal transient simulations can be performed in the time domain. Power transistors operate at high electric fields and usually also at elevated temperatures. The simulations take self-heating into account by coupling the thermal conduction equation to the drift-diffusion equations. Furthermore, SiC and GaN transistors suffer from gate leakage currents and unwanted carrier traps due to material imperfections. Even if the carrier traps have only a small effect on the DC characteristics of a transistor, they usually have a crucial effect on the large signal performance. These effects can be studied in gate lag and drain lag measurements. While drain lag is usually caused by substrate traps gate lag is attributed to surface traps.

List of talks    Main Page

---

Abstract: Contemporary microwave engineering design is heavily based on EM simulations. Simulation-driven design is a must for growing number of devices and systems for which theoretical (e.g., analytical) models are either not available or not sufficiently accurate to yield the design satisfying given performance requirements. Unfortunately, accurate numerical evaluation may be computationally expensive, particularly for complex structures. This makes straightforward approaches, such as employing a simulator directly in the optimization loop, impractical.
Computationally efficient EM-driven design optimization can be realized using physically-based surrogate models. More specifically, optimization of the CPU-intensive structure (high-fidelity or fine model) is replaced by iterative updating and re-optimization of computationally cheap low-fidelity (or coarse) model. The most successful approaches of this kind in microwave engineering include space mapping and simulation-based tuning.
In this talk, a few alternative surrogate-based techniques for microwave design optimization are discussed that are based on response correction of the coarse model. These methods include adaptive response correction, manifold mapping and shape-preserving response prediction. All of these techniques are easy to implement, they do not require extraction of surrogate model parameters (typical for space mapping), nor any modification of the structure of interest (typical for tuning). Also, they can be extremely efficient in terms of yielding a satisfactory design at a low computational cost of a few EM simulations of the optimized structure. Application examples are provided.

List of talks    Main Page

---

Abstract: In recent years requirements for the antenna combiner networks used in base stations for radio mobile telecommunication have undergone a twofold path:
- Selectivity and insertion loss requirement have become more and more stringent
- Complexity (in terms of number of filters and ports) of such devices has increased beyond the typical duplexer structure
This is caused by the increasing need for capacity in restricted spectral bandwidth and results in several cellular communication system sharing the same site, including where possible, feeder cables and antennas (this being driven by economical and environmental considerations as well). To cope with this, a number of new filter-based combiner architectures have emerged.
Aim of this talk is to present a new design approach to modern combiner systems where the design of the filtering network is consider as a whole (rather than to synthesize separately the filters composing the combiner, connect them together and optimize the resulting network); in this way greater flexibility can be achieved in the combiner realization and the overall time requested for going from the design to the fabricated product can be significantly reduced.
The goal of the novel design approach is the evaluation of the characteristic polynomials of the composing filters, taking into account the constraints represented by the interconnections and by the selectivity requirements. Once these polynomials are available, the synthesis of a low-pass prototype of the overall combiner can be carried out with the well known techniques available in the literature; then the dimensioning of the physical structure implementing the synthesized network is accomplished by imposing the coupling coefficients and the external Qs evaluated from the prototype.
The design of two specific types of combiner will be illustrated in the presentation. The first type is constituted by the classical star-junction multiplexers, employing a resonating junction; an iterative technique for the evaluation of the characteristic polynomials of both the overall multiplexer and of the composing filters will be presented and discussed. The second type of combiner is the so called Mast Head Amplifier Combiner, a passive network employing three filters and two junctions (details about its structure and use will be given in the presentation); the evaluation of the characteristic polynomials of this combiner is carried out by means of numerical optimization, adopting suitable frequency transformations to increase the accuracy of the computations.
Several examples illustrating the proposed design procedures will be presented, together with the comparison between the expected behaviour of the synthesised networks and the response obtained by manufactured prototypes.

List of talks    Main Page

---

Abstract: Modern microwave and millimeter-wave equipment, present in mobile, wireless and space communication systems, employ a wide variety of waveguide components. Most of these components are based on the cascade connection of waveguides with different cross-section. Therefore, the full-wave modal analysis of such structures has received a considerable attention from the microwave community.
Modal methods such as mode-matching and integral equation are widely employed in the analysis of planar junctions. Then segmentation techniques are applied, which consist of decomposing the analysis of a complete waveguide structure into the characterization of its elementary key-building blocks, i.e. planar junctions and uniform waveguides. Even though such methods can provide very accurate results in short CPU times, a common drawback is that the generalized matrices must be completely recomputed at each frequency point.
The objective of the proposed talk will be to describe a new method for the analysis of passive waveguide components, composed of the cascade connection of planar junctions. This new method extracts the main computations out of the frequency loop thus reducing the overall CPU effort for solving the frequency-domain problem. The key points to reach such objectives are:
• Starting from the integral equation technique for the representation of planar waveguide junctions, we propose a novel formulation of the generalized impedance and admittance matrices in form of quasi-static terms and a pole expansion. A convergence study of this algorithm will be presented, where the two formulations in form of admittance and impedance matrices are compared in terms of efficiency and robustness. 
• Once the generalized matrices of planar junctions are expressed in form of pole expansions, a technique that provides the wideband generalized impedance or admittance matrix representation of the whole structure in the same form will be presented. For this purpose, the structure is segmented into planar junctions and uniform waveguide sections, which are characterized in terms of wideband impedance matrices. Then, an efficient iterative algorithm for combining such matrices, and finally providing the wideband generalized impedance matrix of the complete structure, is followed. Special formulation will be derived for two-dimension structures in order to obtain more optimized algorithms for this kind of structures widely employed in practical designs.
Finally, the proposed method will be validated through the presentation of several practical designs. The results provided by our method will be compared with the results provided by the previous methods commonly employed for the analysis of such passive devices, and with the results provided by commercial software.

List of talks    Main Page

---

Abstract: Electromagnetic simulations are often used as forward models in the solution of inverse imaging and detection problems. In the microwave frequency range, high-fidelity full-wave simulations are required, which are very time-consuming. This has prevented real-time detection and imaging algorithms from making use of such simulations. Simulation-based forward models have been applied to iterative solutions only, which are known to be prohibitively long and often unreliable. We present one recent development, which enables electromagnetics-based real-time detection and imaging with microwave measurements. The objective is the detection of targets (or scatterers) embedded in an otherwise known background medium. A novel detection approach based on self-adjoint response sensitivity analysis is outlined and illustrated through examples. The sensitivity analysis uses simulated field distributions in the background medium. Since this medium is common for all objects under test, it is simulated only once before the actual tests are performed. Thus detection can be performed in real time.

List of talks    Main Page

---

Abstract: Design of ultra-wideband (UWB) planar antennas is a challenging task due to the lack of good theoretical models and high computational cost of electromagnetic (EM) simulation of the antenna structures. In addition there typically are a few UWB antenna figures of interest which have to meet the design requirements such as reflection coefficient, radiation pattern, introduced phase distortion, etc, all over the UWB band of interest. 
Direct EM-based optimization of UWB antennas is impractical not only because of being CPU-intensive but also because it often fails due to poor analytical properties of the EM-based objective function. Many existing approaches exploit methods like genetic algorithms or particle swarm optimizers which are characterized by huge computational overhead. On another hand, a design practice is multiple simulations and parameter sweeps, where the design parameters are iteratively modified based on experience of the engineer.
Here, rapid design optimization procedures for UWB planar antennas are discussed. The optimization burden is shifted to a surrogate model, computationally cheap representation of the structure being optimized. Presented methods include space mapping with coarsely-discretized EM models, variable-fidelity multi-model algorithm, and surrogate-based optimization exploiting Cauchy approximation. Considered approaches yield reliable designs at computational costs corresponding to a few full-wave simulations of the antenna in question.

List of talks    Main Page

---

Abstract: The “tuning” methodology is illustrated using filter design examples. In the tuning methodology, additional perfectly calibrated ports (“tuning ports”) are inserted in and between filter resonators prior to EM analysis of the layout. Thus, instead of analyzing a 2-port filter repeatedly while trying to adjust the layout for desired performance, a 20 or 30 or 40+ port layout is EM analyzed once. Then tuning elements are connected into the filter by circuit theory (i.e., nodal analysis) connection to the tuning ports. The tuning elements can be circuit theory based, or EM based, or a combination. The net result is that the filter can now be tuned at circuit theory speed with full EM accuracy. For full application, this technique requires perfectly calibrated ports. However, it can actually be realized at lower frequency (where port calibration is not needed) using nearly any EM analysis tool. This technique can be immediately applied using any microwave design framework (e.g., Agilent, AWR, Cadence) that allows interoperability with an appropriate EM analysis tool. Design closure typically decreases from about 2 weeks to 1 day in practice.

List of talks    Main Page

---

Abstract: High Impedance Surfaces (HISs) exhibit unique properties like the in-phase reflection of incident waves and the suppression of the surface waves. Different antenna parameters like the gain, impedance and the size can be enhanced by incorporating the HISs into the antenna structures. The design of the HISs can be optimized to tailor their electromagnetic properties depending on the operational requirements. The HIS is realized by printing the periodic array of Frequency Selective Surfaces (FSSs) over a metal backed dielectric substrate. The design of the HIS presents the difficulties in its optimization because of its size and the resources required. It is observed that the characteristics of the FSS and HIS follow each other and this correlation can be exploited to speed up the optimization process by carrying out the design in two steps: 1. Optimize the free-standing FSS, and 2. Realize the HIS by printing the optimized FSS on a metal backed dielectric substrate. Commercial software FEKO is used for the analysis of the FSS structure using two different optimization methods: Simplex Nelder-Mead and the Particle Swarm Optimization (PSO). Being a local optimizer, the convergence of the Simplex algorithm is much faster compared to the global optimizer PSO. But, unlike the global optimizer PSO, the success of the Simplex depends on the starting point that carries the disadvantage of converging at a local minimum. To improve the chances of reaching the global minimum without compromising on the speed of convergence, PSO is hybridized with Simplex where the global optimizer is used to find the starting point for the local optimizer. The hybridized method has improved the speed of the optimizer without compromising on the ability to reach the global goal. Even though there is growing demand for wideband antennas, a narrowband design has its advantages in cordless phone applications, that operate at 49 MHz, 900 MHz, 2.4 GHz and 5.8 GHz. Interference from adjacent frequency bands can be avoided by using narrowband antenna that will operate only around the frequency of interest. This paper presents a 5.8 GHz low-profile monopole (quarter wavelength) antenna with a 0.07 λ_dielectric thickness, realized by using the narrowband HIS as the substrate. The FSS structure was designed by combining the Jerusalem cross and the three step fractal patch in a single unit cell and the proposed HIS was realized by printing the FSS structure that is optimized for steep behavior of the reflection coefficient, over a metal backed dielectric substrate (RT/duroid 5880 laminate with dielectric constant of 2.2). Numerical results and comparison of different optimization techniques will be presented at the workshop.

List of talks    Main Page

---

Abstract: The purpose of the presentation is to demonstrate a general and rigorous CAD procedure that can potentially provide a systematic answer to the needs of modern circuit-level microwave system simulation by combining nonlinear, electromagnetic (EM), and propagation analysis tools. First the transmitter front-end is analysed under digitally modulated IF drive as a nonlinear system with a load consisting of the antenna described by full-wave EM analysis across the frequency band of interest. In order to produce the radiated field envelope, the EM results are interfaced with an envelope-oriented harmonic-balance technique (MHB) based on Krylov-subspace model-order reduction. The far-field radiated by the antenna, possibly in the presence of inhomogeneous media, is used as the input to an advanced 3D Ray Tracing (RT)-based radio channel model, to compute the channel transfer function in realistic propagation conditions. On output, the reciprocity theorem is applied to the multiport receiving antenna under plane-wave incidence to derive a rigorous circuit description of the receiver excitation. The circuit-level nonlinear analysis of the receiver treated as a whole is again performed by MHB. Possible applications include RFID and UWB systems. The extension to multiple transmitting and receiving antennas is also possible, thus providing for the first time a sys-tematic approach to a full CAD-based treatment of MIMO systems.

List of talks    Main Page

---

Abstract: Design, modeling and optimization of electromagnetic structures electronic devices, circuits and systems, require the application of advanced tools in computational electromagnetics and multiphysics modeling. Especially in the case of nanoelecronic devices in addition to electromagnetic modeling also the consideration of mechanical, thermal, acoustic and quantum mechanical effects and the combination of these models may be required. Tools for multi-physics modeling of interacting phenomena in complex structures are required to meet with the challenges in design of advanced complex nanoelectric and nanoelectronic structures. Special problems arise from the need of self-consistent modeling of interacting physical phenomena under consideration of large variations of space- and time scales, the time variations of geometry and material parameters and the demand for compact model generation and design optimization.
Network-oriented methods applied to field problems may contribute significantly to the problem formulation and solution methodology. Whereas in field theory the threedimensional geometric structure of the electromagnetic field has to be considered, a network model exhibits a plain topological structure. In network theory systematic approaches for circuit analysis are based on the separation of the circuit into the connection circuit and the circuit elements. The connection circuit represents the topological structure of the circuit and contains only interconnects, including ideal transformers.
Applying a network description electromagnetic and multiphysics structures can be segmented into substructures. These substructures define the circuit elements and the set of boundary surfaces between the substructures define the interconnection network. Canonical Foster equivalent circuits can represent lossless structures in sub-domains. Canonical Cauer networks can describe radiation modes. The lumped element models can be obtained by analytic methods, i.e. via Green's function or mode matching approaches or by numerical methods techniques (Transmission Line
Matrix Method or Transverse Wave Formulation) in connection with system identification techniques.
By applying time discretization using Richards transformation a time-discrete transmission line segment circuit (TLSC) algorithm for efficient time-domain modeling of electromagnetic structures is formulated. The TLM scheme is a special case of the TLSC scheme and can be easily incorporated into the TLSC scheme yielding a powerful hybrid method. The application of wave digital filter (WDF) methods for time-discrete modeling and their relation to TLSC and TLM schemes is discussed. The network approach allows a systematic introduction of hybrid methods. Furthermore, network formulations are well suited for the application of model order reduction methods.

List of talks    Main Page

---

Abstract: We analyze the electromagnetic eigenvalue problem arising from a two-dimensional model of dielectric waveguide structures. Such so-called board-integrated waveguides are used to guide optical waves (wavelength around 1310nm) within a classical electrical circuit board (electrooptical circuit board, EOCB). Their transversal size is roughly 300x300 microns with a gradient index profile. From various processing steps, the waveguides show some variations from their design-value in this profile which are supposed to have considerable influence on important properties such as the field pattern of the basic mode, its effective index of refraction (including damping) and the coupling efficiency to standard fibers. The aim of this project is to calculate the sensitivity of such characteristic properties with respect to the material distribution in the cross section.
We use a computational model arising from a discretization with the finite integration technique (FIT) on a two-dimensional Cartesian mesh with N=Nx*Ny grid points. The permittivities eps_i in each mesh cell (i=1...N) are assembled in a diagonal matrix operator M_eps, where some averaging procedures between the cells have to be applied. This finally leads to a sparse, non-symmetric eigenvalue problem, with the (squared) propagation constant beta^2 of the modes as eigenvalue and the discrete electric field as eigenvector.
For small problem sizes and low-dimensional approximations of the parameter space the sensitivities (d beta/d eps_i) can easily be found by extensive parameter studies. However, we also aim at calculating the sensitivities by analytical derivation from the algebraic eigenvalue formulation, using algorithmic differentiation in combination with known results for the computation of derivatives of eigenvalues and -vectors with respect to a parameter-dependent matrix. This approach allows the computation of the derivatives within working accuracy. The presentation will show results of both approaches and compare them in terms of accuracy and efficiency.

List of talks    Main Page

---

Abstract: The talk will present recent advances on stability-analysis methods, applied to power amplifiers, which often exhibit undesired behavior. Techniques will be provided to detect the most common types of bifurcations, or qualitative changes of the solution stability when a parameter is varied continuously. The bifurcation analysis techniques are combined with pole-zero identification in order to detect an undesired coexistence of stable solutions. Bifurcation loci provide instability contours versus practical design parameters and can be used to evaluate the impact of the instability on the overall amplifier performance and to stabilize the amplifier. Various examples of usual instability phenomena in power amplifiers will be presented.

List of talks    Main Page

---

Abstract: Due to the tremendous improvements in numerical techniques and computer hardware, simulations and even optimizations of complex components and systems become more and more practical. Despite the advances in computing resources, the usage of highly specialized simulation techniques is still essential for an efficient handling of such problems. This presentation will show some of the trends in cluster computing and GPU based acceleration techniques. In addition, the presentation will illustrate how a divide and conquer approach including a coupling of various analysis techniques can be used to address large and complex system simulations. For each component within a system, the best suited simulation technique can be applied delivering the best possible efficiency. In addition, this "complete technology" framework can also be used to support very advanced and specialized design tools such as space mapping optimization.

List of talks    Main Page

---

Abstract: The performance of the most commonly used metal antennas close to the human body is one of the limiting factors of the performance of bio-sensors and wireless body area networks (WBAN). Due to the high dielectric and conductivity contrast with respect to most parts of the human body (blood, skin, ...), the range of most of the wireless sensors operating in RF and microwave frequencies is limited to 1-2 cm when attached to the body. In this contribution, we discuss the design and optimization process for this novel and challenging idea, that is based on engineering the properties of liquids. Various 2D and 3D structures and analyzed using numerical tools and practical bio-phantom topologies for operability up to 3 GHz. The liquid antenna topologies  allow for the improvement of the range by a factor of 5-10 and could potentially set the foundation of liquid RF electronics for implantable devices and wearable real-time bio-signal monitoring, since it can potentially lead to very flexible antenna and electronic configurations.

List of talks    Main Page

---

Abstract: In this talk, inkjet-printed flexible antennas, RF electronics and sensors fabricated on paper and other polymer (e.g.LCP) substrates are introduced as a system-level solution for ultra-low-cost mass production of UHF Radio Frequency Identification (RFID) Tags and Wireless Sensor Nodes (WSN) in an approach that could be easily extended to other microwave and wireless applications. The talk will cover examples from UHF up to the millimeter-wave frequency ranges. A compact inkjet-printed UHF “passive-RFID” antenna using the classic T-match approach and designed to match IC’s complex impedance, is presented as a the first demonstrating prototype for this technology. Then, Prof. Tentzeris will briefly touch up the state-of-the-art area of fully-integrated wireless sensor modules on paper or flexible LCP and show the first ever 2D sensor integration with an RFID tag module on paper, as well as numerous 3D multilayer paper-based and LCP-based RF/microwave structures, that could potentially set the foundation for the truly convergent wireless sensor ad-hoc networks of the future with enhanced cognitive intelligence and "rugged" packaging. Prof. Tentzeris will discuss issues concerning the power sources of "near-perpetual" RF modules, including flexible minaturized batteries as well as power-scavenging approaches involving thermal, EM, vibration and solar energy forms. The final step of the presentation will involve examples from wearable (e.g. biomonitoring) antennas and RF modules, as well as the first examples of the integration of inkjet-printed nanotechnology-based (e.g.CNT) sensors on paper and organic substrates. It has to be noted that the talk will review and present challenges for inkjet-printed organic active and nonlinear devices as well as future directions in the area of environmentally-friendly ("green") RF electronics and "smart-skin' conformal sensors.

List of talks    Main Page

---

Abstract: Recent advances in the application of Artificial Neural Networks (ANN) to radio-frequency (RF) and microwave design created an exciting direction of computer-aided modeling and design of high-frequency electronics. ANNs are trained to learn the high-frequency behavior of electronic components, and trained ANNs can be used as models for high-level electronic design.  The ANN models are much faster than detailed electromagnetic/physics based models of electronic components, and more accurate than conventional empirical/equivalent circuit models. It leads to substantial increase in modeling accuracy, speed, and flexibility. Applications are being made in modeling and design of passive and active RF/microwave electronic components and circuits, high-speed VLSI interconnects, printed antennas, LTCC circuits, semiconductor devices, measurement standards, filters, amplifiers, mixers and so on.  Automated model generation algorithms integrating data generation and ANN training are being developed. Knowledge based neural networks exploiting prior knowledge such as empirical/semi-analytical models are being introduced in microwave computer-aided design (CAD). This leads to new level of CAD methodologies combining equivalent circuit/empirical models, electromagnetic/physics simulation and behavioral modeling with ANN and optimization algorithms for fast and accurate design of high-frequency circuits and systems. This talk presents a review of the state of the art in these emerging directions.  The presentations highlight implementable methodologies for automated modeling and design of high-frequency electronic components, circuits and systems.  The presentation covers fundamental concepts and methodologies, industrial applications, and future trends in R&D.

List of talks    Main Page

---