1,721,047 research outputs found
A Case Study for Improving Performance of Frequency Selective Surface through Union of Sub-Sets and Particle Swarm Optimization
Get PDFFrequency Selective Surfaces (FSSs) consist of a repetition of a given pattern in a periodic way; typically, a dielectric substrate supports this arrangement giving rise to a two-dimensional array. Although relatively simple in structure, designing an FSS that exhibits large bandwidth and stable response to oblique incidence is not straightforward and requires special attention and significant computational effort. To address this problem, this study presents a methodology whereby an initial configuration of the FSS pattern is subjected to an optimization method for sizing the geometrical parameters. Consequently, the initial unit cell is first broken down into subsections, specifically as a “union of subsets”, then particle swarm optimization is used to achieve optimal design parameters that further improves the overall FSS performances. To validate the proposed method, an X-band FSS is proposed and optimized in a commercial simulation environment (Microwave Studio, Dassault Systèmes)
Fully Automated Inside Body WDT Transmitter Design and Optimization Through Artificial Intelligence-Based GANs and DNNs
Get PDFBiomedical inside body wireless data transfer (WDT) interface includes the design of power amplifiers (PAs) with implantable antenna leading to operate concurrently. Hence, active and passive devices are utilized simultaneously for which the accurate starting points for designing these high dimensional devices is critical. From another point of view, accelerating the design and optimization process is another substantial issue that must be considered effectively. In this study, we propose a methodology that includes two optimization phases that are applied sequentially. In the first phase, the PA is designed and optimized by employing a generative adversarial network (GAN) for predicting the load-pull contours on the Smith chart and using a long short-term memory (LSTM)-based deep neural network (DNN) for achieving the optimal design parameters of the biomedical amplifier. In this step, the GAN leads to predicting the optimal impedances needed to construct the initial structure of PA through a simplified real frequency technique. In the second optimization phase, the initial structure of the biomedical antenna is constructed automatically by developing a visual basic (VBA) environment, then like the PA, the design parameters of the antenna are optimized through the LSTM-based DNN. Finally, another GAN is generated for predicting the radiation patterns of the antenna. In both phases, a multi-objective ant lion optimizer (MOALO) is employed in the output layer of DNNs for optimizing various outcome specifications. The proposed method is performed fully automatically: active and passive devices are designed and optimized with the help of GANs and DNNs in which the drawback of heavy reliance of the system performance on the designer's experience is solved in a fast way. The proposed method is validated by designing and optimizing a biomedical PA with an antenna working at the center frequency of 2.45 GHz which shows reliable outcomes
Surrogate Modeling for Designing and Optimizing MIMO Antennas
Get PDFThis papers presents the design and optimization of multiple-input and multiple-output (MIMO) antennas through intelligent methods namely as: surrogate modeling. The optimization process is performed automatically with the combination of Microwave Studio (Dassault Systèmes) and MATLAB numerical analyzer. The proposed optimization method aims to find the optimal solution for the total active reflection coefficient (TARC) specification, S 11 , and S 12 by using shallow neural network. This methodology leads to efficiently size the design parameters of MIMO antenna and to optimize S-parameters and TARC specification jointly. To validate the proposed method, an ultra wideband MIMO antenna in the frequency band of 3.1 GHz to 10.6 GHz is designed and optimized
Nonlinear Behavioral Modeling of FETs: Toward the Implementation of Deep Neural Networks Through Large Signal Data and EDA Tools
Get PDFNonlinear behavioral Field-Effect Transistor (FET) models often rely on large look-up tables extracted from extensive load-pull characterization. Besides the numerical burden, these models have limited extrapolation capabilities and can hardly be made dependent on the device technology. In this paper, we demonstrate that a Long Short-Term Memory (LSTM)-based Deep Neural Network (DNN) is an effective alternative modeling approach. The DNN is trained with the load-pull data within a simulation platform where data exchange between an Electronic Design Automation (EDA) tool (such as PathWave ADS) and a programming platform (such as MATLAB) is exploited. As a test case, the DNN model has been extracted for an S-band MACOM 10W GaN power device, for which the Enhanced Poly-Harmonic Distortion (EPHD) behavioral model is also available in the ADS. The accuracy of DNN model is verified against the EPHD model in terms of output power, gain, efficiency, and dynamic load lines. Compared to other behavioral models, the DNN approach is expected to provide superior extrapolation capability and to be easily reconfigurable to add/combine heterogeneous device data e.g. from advanced characterization, including memory, and physical (TCAD, EM) simulations
Combinational of surrogate modeling and particle swarm optimization for improving the electromagnetic performances of a frequency selective surface
Get PDFFrequency-selective surfaces (FSSs) consist of the repetition of unit cells for controlling reflection, transmission/absorption of electromagnetic (EM) fields. They are typically employed at radio and optical frequencies. Simulation of such large (in terms of wavelength) structures based on the traditional EM simulations is time-consuming and requires significant computational resources. Hence, this paper devotes to present an optimization-oriented methodology for designing and optimizing FSS in an automated fashion. The FSS structure is optimized using the artificial neural network paradigm, where the particle swarm optimization is applied for sizing the design parameters. The optimization process is an automatic one where electronic design automation tool with numerical analyser is working together, leading to effectively optimize the FSS design. To verify the effectiveness of the proposed method, an FSS structure exhibiting a wide transmission band for normal incidence in the 7.0-11.2 GHz range is considered
Deep neural learning based optimization for automated high performance antenna designs
Get PDFThe present paper introduces an optimization-oriented method here practiced for designing high performance single antennas in a fully automated environment. The proposed method comprises two sequential major steps. The first one devotes configuring the shape of antenna and determining the feeding point by employing the bottom-up optimization (BUO) method. In this algorithm, the number of microstrip transmission lines (TLs) used to model the radiator is increased consecutively and the shape of the antenna is revised up to finding the initial satisfying results. Secondly, for determining the best design parameters of the configured antenna shape in the first step (i.e., width and length of TLs), deep neural network (DNN) that is based on Thompson sampling efficient multi-objective optimization (TSEMO) is applied. The recommended optimization method is successfully attracted as a problem solver for designers to tackle the subject for antenna design such as the complexity and large dimensions of structures. Hence, the main advantage of the implemented optimization method in this article is to noticeably decrease the required designer’s involvement automatically generating valid layouts. For validating the suggested method, two wideband antennas are designed, prototyped and subjected to experiment. The first optimized antenna covers the frequency band 8.8–10.1 GHz (13.75 % bandwidth) characterized by a maximum gain of 7.13 dB while the second one covers the frequency band 11.3–13.16 GHz (15.2 %) which exhibits a maximum gain of 7.8 dB
Prediction of Class-Amplifiers with the Aid of Neural Network
Get PDFThis paper presents a strategy addressing the problem of selection of the class of the amplifiers to be used in future wireless communication systems. The proposed methodology uses a scheme based on neural networks (NN): the characteristics of each class of amplifier (i.e., A, B, AB, C, D, F, G, J, S, T , etc.) are determined and then the ‘classification NN’ is constructed for distinguishing various classes from each other. To validate the method, firstly the designs of various class-amplifiers are collected from the recently published literature, and then the specifications of the amplifiers are extracted in terms of voltage (V), current (I) and efficiency; finally with these data the classification NN is trained. After building this black-box NN, providing the required specifications of each amplifier, designer are informed about the class of amplifier that is predicated by the classification NN and that better fits the characteristics of the considered application. This methodology is important as it leads the way of amplifier class selection in the complex communication systems
Combinational of GAN and LSTM-Based DNN for Automatic Optimization of Active and Passive Devices
Get PDFFuture high-data-rate communication systems include concurrent utilization of active and passive devices as power amplifiers (PA) and antennae. Hence for these high dimensional designs, determining an accurate starting point along with achieving acceptable outcomes are required effectively. This paper is devoted to presenting the implementation of two types of neural networks: generative adversarial network (GAN) and long short-term memory (LSTM) deep neural networks (DNNs) for both PA and antenna devices. The benefit of implementing GAN for the PA side is to estimate the load-pull contours on the Smith chart as optimal gate and drain impedances. In addition, the GAN is employed on the antenna side to predict the radiation pattern outcomes for the determined frequency. From another point of view, the LSTM-based DNN with the utilization of the Thompson sampling efficient multi-objective optimization (TSEMO) is presented for predicting the optimal design parameters leading to achieving the targeted specifications for both active and passive devices. The presented methodology is validated by designing and optimizing a PA with a multipleinput and multiple-output (MIMO) antenna operating at an approximate bandwidth of 2.68 GHz
Performance Prediction of Power Amplifiers for the Extended Bandwidth via Neural Networks
Get PDFThis paper presents the optimization methodology
for modeling the power amplifier (PA) with the aid of deep neural
network (DNN). In this paper we propose an impressive approach
leading to extrapolate frequency responses of the PA, where
the long short-term memory (LSTM) DNN is employed. The
presented method models the PA accurately in terms of scattering
parameters, gain, output power and efficiency. This approach
tackles the problem of dependency to the engineer experience
and reduces the challenges in achieving large frequency band.
All the modeling process is performed with the combination of
electronic design automation tool and numerical analyzer where
automated environment is created. For validating the proposed
method, one PA is designed and modelled for the range frequency
of 1 to 2.3 GHz. The DNN is firstly trained for the half of the
bandwidth and later, the modeled PA is used for predicting the
extended frequency band
Beam-Steering Antenna Technique Using Operational Amplifiers for Sub-6 GHz
Get PDFA methodology for beam-steering using operational amplifiers (Op-amps) is presented. Continuous steering is required in various advanced applications and its implementation necessitates additional efforts since singularly control of the feeding of the radiators is mandatory for both phase and amplitude. The present work proposes a technique for generating the required sequence of the feeding signal using two Op-amps for each input port. It leads to generating incremental phases with sequential Op-amps without any limitation in the value of the phase differences, controlled by the bias voltage applied to the Op-amps, hence giving rise to a continuous beam-steering capability. The study case consists of a four-stage oscillator designed for creating a sequence of signals with progressive phase shifts between consecutive outputs. The general scheme allows continuous control of the phase differences here applied for generating a uniform, i.e., constant signal magnitude, feeding sequence. This set of signals is then used to feed a four-element microstrip array operating at 1.2 GHz. The effectiveness of the method is validated by numerical simulation of the array performances. Additionally, the low power consumption of active Op-amps, easy implementation, and high sensitivity are characteristics of the presented paradigm
- …
