1,720,981 research outputs found
Hyperparameter Optimization of Long Short-Term Memory Based Forecasting DNN for Antenna Modeling through Stochastic Methods
Full text linkThis letter presents an impressive optimization method for determining the optimal model hyperparameters of a deep neural network (DNN) targeted to model the characteristics of antennas. In this paper we propose an innovative approach of efficient yield analysis for modeling and sizing antennas. It is based on the long short-term memory (LSTM) DNN aiming to forecast the extended frequency responses, where various stochastic methods are applied for determining the optimal hyperparameters while training a DNN. Among the various methods, the one which models the antenna accurately in terms of input scattering parameter, gain, and radiation patterns is the winner. The proposed method is compact and addresses the problem of heavy reliance to the designer experience in determining the hyperparameters. Additionally, forecasting the future frequency responses of the antenna reduces the designers effort substantially in measuring large frequency band; hence, measuring whole frequency band would not be needed. For validating the effectiveness of the proposed method, the fabricated two element antenna array is used for modeling where the results demonstrate that the Thompson sampling (TS) algorithm can determine optimal hyperparameters with minimum error in comparison with other reported stochastic methods leads to predict the future frequency band accurately. IEE
The Magic of Quantum Computing for Microwave Computer-Aided Design: A Brief Overview
Full text linkQuantum computing has recently become an effective technology for tackling the drawbacks of microwave links where these were previously almost impossible. In that context, the quantum technology has made it possible not only to design a compact size system with improved performance but also to reduce the computational power and time. This presentation aims to provide a short overview through cases studies of the recent use of quantum computing in the design of the elements employed in the receiving section of a transceiver, such as active low-noise amplifier and passive phased-array antenna. This paper encourages the future studies in the domain of quantization effect of electromagnetic energy opening the way for other applications such as quantum communications and/or quantum radar schemes
Multi-band Implantable Microstrip Antenna on Large Ground Plane and TiO2 Substrate
Full text linkBiomedical implanted devices are typically used for interacting with organs and/or for investigating various physiological signals. Hence, enhanced performance devices for clinical uses have got the attention of researchers. In this study, a multi-band implanted microstrip antenna suitable for transmitting/receiving biomedical signals in the Industrial, Scientific and Medical (ISM) frequency bands is presented. The antenna is built on a bio-compatible substrate, as titanium dioxide (TiO2) with relative permittivity of 95. The ground plane is thought to be a bio-metallic implant located within a bone. The proposed antenna is compact in size, 14 × 18 × 1.6 mm3, and works in both 2.45 GHz and 5.8 GHz centered frequency bands. It is designed and optimized considering the actual biological tissues as bone, muscle, fat, and skin surroundings. The simulation results referring to a planar stratification prove that the multiband single microstrip antenna is working properly within the human body and it can be used for medical communication services
Tunable Frequency Selective Surface Design Using Automated Random Optimization
Full text linkWe present an automated approach to design a high performance, tunable frequency selective surface (FSS). The main goal of this study is to provide the simultaneous optimization of the FSS structure in two states of the 4 incorporated varactors, aiming to get an acceptable polarization filtering and polarization control. Generally, microwave designs are dealing with a large amount of data and they depend on the engineer's experiences. In order to get rid of this dependency and providing a ready-to-fabricate layout, we propose an optimization-oriented method based on the random optimization (RO). The RO method is applied in an automated environment where HFSS and Matlab are collaborating together forming a co-simulation platform where the design parameters are optimized up to achieve suitable output performances
Multi-objective Optimization Methods for Passive and Active Devices in mm-Wave 5G Networks
Full text linkDue to the exponential growth of data communications, millimeter-wave (mm-Wave) new radio specification becomes key enablers for fifth generation (5G) communication systems. However in the mm-Wave band frequency, the propagation loss is intensively large and cannot cover all the determined specifications. To tackle this drawback, the transceiver parts must sense the high radiated output power from power amplifiers. Hence by using high performance wideband antennas, the amplifiers can facilitate massive multiple-input multiple-output (MIMO) 5G systems. The figure of merit (FoM) of an amplifier is determined by the output power that must be challenged by other design specifications as: power gain, drain efficiency, and linearity. Therefore, powerful multi-objective optimization methods are required for welcoming appointed passive (antennas) and active (power amplifiers) characteristics in the determined frequency band. On the other side, high performance antennas in the 5G networks are also needed that can be designed using potent optimization methods. In this chapter, we provide collection of various optimization methods which have been recently applied for designing and optimizing high performance high power amplifiers and antennas. Hence, any designer can access to the nominated algorithms and can select the ones that are suitable for their problems
Modeling of Biomedical Antennas through Forecasting DNN for the Enlarged Bandwidth
No full textRecently, wireless medical technologies are growing day-by-day resulting in complex structures and topologies. Hence, advanced methods are required for designing and optimizing biomedical devices subject to high-dimensional parameter space. This paper is devoted to presenting an effective approach for estimating frequency responses of an implanted, multiple-input multiple-output (MIMO) antenna through the deep neural network (DNN) in terms of S11, S12, and total active reflection coefficient (TARC) specifications. This impressive approach aims to facilitate the time-consuming simulations in large multi-frequency bands and concurrently reduce the dependency on the designer's experience. All the process is performed in an automated environment and the proposed method is verified by designing and optimizing an implanted MIMO antenna operating in frequency bands of 4.34-4.61 GHz, and 5.86-6.64 GHz. In this design, the Long Short-Term Memory (LSTM)-based DNN is trained for the frequency band between 3-5.8 GHz, and afterward the constructed DNN is employed for predicting the various antenna specifications for the future bandwidth of 5.8-8 GHz
Optimization for wideband linear array antenna through bottom-up method
Full text linkThis paper presents an automated design methodology for electromagnetic- based (EM-based) optimization of an array antenna by applying bottom-up approach. Firstly, one single antenna is optimized then bottom-up optimization (BUO) method has been implemented by increasing the number of single antennas, sequentially. The proposed method leads to automatically find an optimal array by setting the distance between single antennas. The optimization method is performed in an automated environment with the help of an electronic design automation (EDA) tool and a numerical analyzer. The results of the final design have been compared by means of two EDA tools such as ADS and HFSS. The optimized array antenna works in the frequency band from 12.9 GHz to 14.3 GHz. It offers a linear gain performance higher than 7.5 dB. The simulations in both ADS and HFSS tools illustrate a good match in S-parameter and gain simulation output results
MIMO Antenna Optimization: From Configuring Structure to Sizing with the aid of Neural Network
Full text linkIn the last decades, multiple input, multiple output (MIMO) antenna designs play important role and this trend will continue in next-generation mobile technologies. Designing high-performance MIMOs is significant since these types of antennas include multiple radiating elements. For these complex configurations, intelligent-based optimization methods can tackle the problem of designing. This paper devotes to designing and optimizing the configuration and design parameters of a MIMO antenna, respectively. Firstly bottom-up optimization (BUO) approach is executed successfully for building the general topology of the MIMO antenna and afterwards, artificial neural network (ANN) is utilized for obtaining the design parameters with the optimal values. The proposed approach results in generating the optimal topology with size values in a reduced effort by designers. The presented approach is applied to designing a MIMO antenna operating from 13.7 GHz to 29 GHz
Automated optimization for broadband flat-gain antenna designs with artificial neural network
Full text linkAn automated optimization process for designing and optimising high-performance single microstrip antennas is presented. It consists of the successive use of two optimization methods, bottom-up optimization (BUO) and Bayesian optimization (BO), which are applied sequentially, resulting in electromagnetic (EM)-based artificial neural network modelling. The BUO method is applied for the initial design of the structure of the antennas whereas the BO approach is successively implemented to predict suitable dimensional parameters, leading to broadband, high flat-gain antennas. The optimization process is performed automatically with the combination of an electronic design automation tool and a numerical analyser. The proposed method is easy to use; it allows one to perform the design with little experience, because both structure modelling and sizing are performed automatically. To verify the power of the proposed EM-based method experimentally, two single microstrip antennas have been designed, optimised, fabricated, and measured. The first antenna has flat-gain performance (6.9–7.2 dB) in a frequency band of 8.8–10 GHz. The second has been designed to perform in the 8.7- to 10-GHz band, where it exhibits flat-gain performance with reduced fluctuation in the range of 6.7–7 dB. The experimental results are in good agreement with the numerical data
MIMO Antenna Design Through Genetic Algorithm
Full text linkMassive multiple-input multi-output (MIMO) is a technology established more than a decade ago; since then it has been employed as an integral part of the fifth generation (5G) mobile technologies. Recently, MIMO antenna designs have taken the attention of researchers due to their benefits such as reduced bit errors, high throughput, low latency and others. MIMO antennas consist of multiple radiating elements leading to have complex designs. Hence, frequently, optimization-based methods are required during their design; this paper provides one possible optimization-based methodology for designing them. The design procedure starts with configuring the MIMO antenna and then employing the genetic algorithm (GA) for optimizing the design parameters describing the MIMO antenna geometry. Using the GA method, leads to have a valid electromagnetic (EM)-verified post-layout generation and reduces the efforts of any designer in providing ready-to-fabricate layout. For validating our proposed method, we design and optimize MIMO antenna with a bandwidth of 400 MHz in the frequency band 4.1 GHz–4.5 GHz
- …
