Search

Background: Thyroid disorders are prevalent worldwide and impact many people. The abnormal growth of cells in the thyroid gland region is very common and even found in healthy people. These abnormal cells can be cancerous or non-cancerous, so early detection of this disease is the only solution for minimizing the death rate or maximizing a patient's survival rate. Traditional techniques to detect cancerous nodules are complex and time-consuming; hence, several imaging algorithms are used to detect the malignant status of thyroid nodules timely. Aim: This research aims to develop computer-aided diagnosis tools for malignant thyroid nodule detection using ultrasound images. This tool will be helpful for doctors and radiologists in the rapid detection of thyroid cancer at its early stages. The individual machine learning models are inferior to medical datasets because the size of medical image datasets is tiny, and there is a vast class imbalance problem. These problems lead to overfitting; hence, accuracy is very poor on the test dataset. Objective: This research proposes ensemble learning models that achieve higher accuracy than individual models. The objective is to design different ensemble models and then utilize benchmarking techniques to select the best model among all trained models. method: This research investigates four recently developed Transformer and Mixer models for thyroid detection. The weighted average ensemble models are introduced, and the model weights are optimized using Hunger Games Search (HGS) optimization algorithm. The recently developed distance correlation CRITIC (D-CRITIC) based TOPSIS method is utilized to rank the models. Methods: This research investigates four recently developed image transformer and mixer models for thyroid detection. The weighted average ensemble models are introduced, and model weights are optimized using the hunger games search (HGS) optimization algorithm. The recently developed distance correlation CRITIC (D-CRITIC) based TOPSIS method is utilized to rank the models. Results: Based on the TOPSIS score, the best model for an 80:20 split is the gMLP+ViT model, which achieved an accuracy of 89.70%, whereas using a 70:30 data split, the gMLP+FNet+Mixer-MLP has achieved the highest accuracy of 82.18% on the publicly available thyroid dataset. Conclusion: This study shows that the proposed ensemble models have better thyroid detection capabilities than individual base models for the imbalanced thyroid ultrasound dataset.

In this paper, the authors propose to design thinned sparse planar array antenna in multiple planes using meta-heuristic optimization algorithm such as Teaching Learning-Based Optimization (TLBO) a...

This study illustrates the dynamical reconfiguration of a concentric hexagonal antenna array radiation to generate a pencil beam and flat-top beam simultaneously by electronic control in two principle vertical planes under consideration. Both the beams share a common normalized optimal current excitation amplitude distribution while the optimal sets of phase excitation coefficients are varied radically across the hexagons to generate a flat-top beam. The proposed approach is able to solve the underlying multi-objective problem and flexible enough to the efficient implementation of additional design constraints in the considered φ-planes. In this paper, a set of simulation-based examples are presented in an integrated way. The outcomes validate the effectiveness of the stated optimization using meta-heuristic optimization algorithms (teaching–learning-based optimization, symbiotic organism search, multi-verse optimization) to reach the solution globally and prove actual relevance to the concerned applications.

This paper presents a study on the synthesis of a position-phase optimized reconfigurable linear array antenna with uniform amplitude distributions. The objective is to produce a pencil/flat-top beam pair. In this position-phase method, a pencil beam is duly generated with zero degree phases and phases are varied between −180° and 180° to produce a flat-top beam keeping position of the elements and amplitude excitations common to both pencil and flat-top beams. The amplitude distribution of the elements is kept uniform. The phases as well as position of the elements are optimized by Teaching Learning Based Optimization (TLBO), modified Quantum Particle Swarm Optimization (QPSO) and Symbiotic Organisms Search (SOS) algorithms to produce the beam pair. The simulations are done for two Sets of elements and a null placement is included in one of the Sets. The results obtained using these algorithms are duly compared with each other and it is found that SOS algorithm performed in par with TLBO algorithm in the generation of the radiation pattern parameters and better over TLBO and QPSO in statistical values.

In this paper, the optimal reduction of side lobe level with constraints of a suitable first null beam width of a non-uniform concentric elliptical array antenna of uniformly excited isotropic elem...

This paper presents a method based on the Moth Flame optimization evolutionary algorithm for the synthesis of Flat-top/pencil beam pair of dual beam reconfigurable concentric circular antenna array...

A compact CPW-fed slot antenna is proposed here. The antenna has a pentagonal patch nested within an asymmetrical slot along with a coplanar waveguide feed. The suggested antenna also has a circula...

This paper proposes the novel application of three meta-heuristic optimization algorithms namely crow search algorithm, moth flame optimization, and symbiotic organism search algorithm for the synthesis of uniformly excited multiple concentric ring array antennas. The objective of this work is to minimize the sidelobe level (SLL) and maximize the peak directivity simultaneously. Three different cases are demonstrated separately with various constraints such as optimal inter-element spacing and/or optimal ring radii. Comparative study of the algorithms using common parameters such as SLL, directivity, first null beam width, best cost, and run time has been reported. Investigation results prove the superiority of case 3 over other cases in terms of directivity and SLL. This work demonstrates the potential of these algorithms.

In this paper, a method relying on Symbiotic Organisms Search optimisation algorithm is proposed for the design of a reconfigurable concentric circular array with phase-only controls differentiatin...

This article deals with the synthesis of uniformly spaced elliptical array antenna using multi-verse optimization algorithm. Two cases are performed here. In case 1 a uniform elliptical array antenna is optimized to obtain a flat top beam pattern with 20 numbers of array elements and a specific semimajor axis value. In case 2, a 40 element elliptical array is designed with optimized semi-major axis. For both the cases the current excitation amplitude is varied within a specific boundary value to get the flat top shape in the radiation pattern. Case 1 mainly focuses on minimizing the ripple in the main beam and case 2 focuses on reducing the side lobes. The challenge of this study is to produce a flat-top pattern with a negligible amount of ripple in the main beam and lower side lobes which can essentially be applied with benefit of constant intensity over a given cross-section. The study shows the robustness of the algorithm and the usefulness of the considered optimization technique.

This study shows the performance analysis of a uniformly excited thinned elliptical array antenna of isotropic antennas using the Teaching Learning Based Optimization (TLBO) algorithm. Here objective is to reduce the side lobe level (SLL) value for a specific thinning percentage. A 40 element elliptical antenna array has been considered and two cases of thinning are performed by targeting different thinning percentages. It is to be noted that with increasing thinning percentage, the value of SLL decreases. The obtained value of the SLL after thinning is better than the SLL of a fully populated antenna array. The proposed approach is simple, economical and practically designable.

In this paper, a novel method for designing unequally spaced time-modulated arrays (UESTMAs) with fewer optimization parameters in two steps has been investigated. Different cases of synthesis method such as position-only (PO), on-time only (OTO), on-time then position (OTTP) and position then on-time (PTOT) have been considered to design the array. For each case under consideration, the global search stochastic algorithm based on differential evolution (DE) has been employed to achieve the array pattern with significantly suppressed sidelobe levels (SLLs) and sideband levels (SBLs) by separately considering the element position and switch on-time of the array elements as the optimization parameters. Through detail comparative results, it is shown that the performance of PTOT synthesis method with fewer optimization parameters in two steps provides better or equivalent performance as compared with the reported approach where both position and on-time are considered simultaneously. The representative numerical results for a 32 element linear array have been presented to show the effectiveness of the proposed approach.

This communication is about new application of a well-known metaheuristic algorithm called Teaching–learning-based optimization towards a concentric circular array antenna for achieving minimized sidelobe levels. The problem under consideration is constrained with inter-ring radii and number of antenna array elements. Elements of the array are isotropic in nature and uniformly excited. Various array antenna parameters like best cost, mean value, standard deviation and computation time are noted in order to summarize the overall antenna performance. The effectiveness of the proposed technique is demonstrated with the help of comparing its result with some benchmarks. Optimizing the number of array elements yields best result with sidelobe levels of nearly −34.5 dB considering 9 concentric circles with common centre as example. Simulations are done using MATLAB with Intel i5 processor.

In this paper, a novel approach to synthesize sum and difference pattern through time modulation in unequally spaced linear arrays with uniform amplitude excitation is presented. Both the patterns are obtained with same element position, whereas the two patterns differ only in switch on-time sequence of the fast electronically controlled switches commutating the radiators. Stochastically, the global exploring capability of differential evolution (DE) is employed to obtain the optimized element position and on-time sequence of the antenna elements. The suitability and usefulness of the proposed method are illustrated through representative results of a 20 element linear array.

Researchers on antenna arrays usually neglect the effect of mutual coupling of antennas placed in proximity to each other. The interchange of electromagnetic energy happening between an antenna and a far field point depends not only on the transmitting antenna, but also from its neighbouring antennas. This effect is referred to as mutual coupling between dipole antenna elements and it is considered here in the synthesis of phase only reconfigurable antenna arrays. The main objective of this work is to produce desired Side Lobe Level and Voltage Standing Wave Ratio in addition to few other radiation pattern parameters. Multiverse Optimization algorithm is employed for the purpose of generating voltage amplitude and discrete phase distributions in the dipole elements for the generation of flat-top beam/pencil beam patterns. These two patterns share the common amplitude distributions and differ in phase distributions. Results obtained using MATLAB simulations prove that this algorithm accomplished its task successfully and is also found to be superior over other algorithms like Particle Swarm Optimization, Grey Wolf Optimization and Imperialist Competitive Optimization algorithms.

This paper describes a method using Quantum Particle Swarm Optimization to obtain a broadband frequency invariant pattern for synthesis of nonuniformly spaced linear array of isotropic antennas. Two cases related to this work using QPSO have been studied, namely, in the first case, the generated frequency invariant far-field pattern is broadside in the vertical plane and in the second case, the far-field frequency invariant pattern is scanned in a particular direction. An effort is made such that the side lobe level and first null beam widths of the patterns are made similar to their related desired values. The two cases are presented in this paper to show the effectiveness of the proposed method in achieving the desired specifications. Even though, the developed method is utilized here for a linear array of isotropic antennas; it can be extended to other type of arrays. The performance of this algorithm is validated by duly comparing it with firefly algorithm.

This paper presents a comparative analysis of three evolutionary algorithms, namely, Backtracking Search Algorithm, Cuckoo Search Algorithm and Artificial Bee Colony Algorithms for synthesis of a scanned linear array of uniformly spaced parallel half wavelength dipole antennas. Here, antenna parameters, namely Side Lobe Level, reflection coefficient and wide null depth are taken into consideration for comparison between algorithms. In addition to it, statistical parameters, namely best fitness value, mean and standard deviation of the fitness values obtained from algorithms are compared. Mutual coupling that exists among the antenna elements is included in obtaining radiation patterns and the self-impedances along with the mutual impedances are calculated by induced Electro-Motive Force method. Two different examples are shown in this paper to validate the effectiveness of the utilized approach. Although, this approach is applied to a linear array of dipole antennas; this can be utilized for other array geometries as well.

A method based on quantum particle swarm optimization is proposed for correcting radiation pattern of a uniformly spaced linear array of parallel half-wavelength vertical very thin dipole antennas when more than one element of the array is completely defective. The generated pattern is broadside in the horizontal plane. Mutual coupling between the half-wavelength parallel dipole antennas has been taken care of by induced EMF method considering the current distribution on each dipole to be sinusoidal. Side lobe level and minimum return loss of the corrected pattern with defective elements are also calculated and an attempt is made that these parameters be equal to their respective specified values. This is usually done by changing excitation current amplitude of the non-defective elements without replacing the faulty elements. The element pattern of half-wavelength dipole antenna has been assumed omni-directional in the horizontal plane. Two examples are presented to show the effectiveness of the proposed approach.

In this paper, a pattern synthesis based on a multiobjective optimization algorithm is proposed for the generation of a reconfigurable pencil/flat top dual-beam planar antenna array built using isotropic antenna elements in selected phi cuts. These beams claim the same amplitude excitations and differ from each other in phase excitations. Zero-phase excitations are used in pencil beam and these phases are updated with optimum phases for the flat top beam. All the excitations are obtained using Moth–flame optimization algorithm. With the support of the fitness functions, care is taken to control the expected values of the radiation pattern parameters to remain under certain fixed limit. In addition, synthesis is also done for the provision of a null in a particular direction for rejection of interference in the pencil beam in two different phi cuts. To suppress the mutual coupling effects, dynamic range ratio is kept under a threshold limit. Simulation results show the effectiveness of this proposed synthesis for phi cut planes. This algorithm is compared and proved to be better in many aspects over the standard meta-heuristic algorithms like Artificial Bee Colony and Imperialist Competitive algorithms in terms of performance parameters.

The authors in this paper present comparative study of three evolutionary algorithms such as quantum particle swarm optimization, backtracking search algorithm and cuckoo search algorithm for design of linear array of dipole antennas that will generate far-field radiation pattern in the horizontal plane with fixed value of peak side lobe level, voltage standing wave ratio and wide null depth. The performances of the above algorithms are judged here in terms of different antenna parameters like peak side lobe level, wide null depth and VSWR as well as statistical parameters such as best fitness value, mean and standard deviation. Mutual coupling between parallel dipole antennas has also been considered. Assuming the current distribution on each thin dipole to be sinusoidal, self and mutual impedances are calculated by induced EMF method. One example is presented to show the effectiveness of the proposed approach.

A compact circularly polarized CPW-fed slot antenna is proposed here. The antenna consists of a pentagonal patch within an asymmetrical slot with a single coplanar waveguide feed. The proposed antenna is excited with CPW feed mechanism and investigated experimentally. The structure is fabricated on FR4 epoxy substrate with a permittivity of 4.4. The impedance bandwidth of 10[Formula: see text]GHz (4[Formula: see text]GHz to 14[Formula: see text]GHz) and the axial ratio bandwidth of 1.2[Formula: see text]GHz with the corresponding fractional bandwidth of 113%. The simulation results fulfil the bandwidth requirements of IEEE 802.11a (5.15–5.35[Formula: see text]GHz/5.47–5.725[Formula: see text]GHz) for wireless applications. In terms of bandwidth, compactness and circular polarization comparable results between simulated and measured results clearly show the validity of the proposed structure.

Thinning a large concentric ring array by an evolutionary algorithm needs to handle a large amount of variables. The computational time to find out the optimum elements set increases with the increase of array size. Moreover, thinning significantly reduces the directivity of the array. In this paper, the authors propose a pattern synthesis method to reduce the peak sidelobe level (peak SLL) while keeping first null beamwidth (FNBW) of the array fixed by thinning the outermost rings of the array based on Gravitational Search Algorithm (GSA). Two different cases have been studied. In the first case only the outermost ring of the array is thinned and in the second case the two outermost rings are thinned. The FNBW of the optimized array is kept equal to or less than that of a fully populated, uniformly excited and 0.5 λ spaced concentric ring array of same number of elements and rings. The directivity of the optimized array for the above two cases are compared with an array optimized by thinning all the rings, while keeping the design criteria same as the above two cases. The optimized array by thinning the outermost rings gives higher directivity over the optimized array by thinning all the rings. Time required for computing the optimum elements state for the above two cases using GSA are shown lesser compared to the optimized array by thinning all the rings using the same algorithm. The peak SLL and the FNBW of the optimized array for the above two cases are also compared with the optimized array by thinning all the rings.

In this paper, the authors have proposed a method based on Quantum Particle Swarm Optimization (QPSO) algorithm in the context of radiation pattern correction of a linear array of isotropic antennas corrupted with one or more faulty antenna elements. Care is taken to maintain the values of side lobe level and maximum Wide Null Placement of the corrected pattern to be identical to the values of non-defective radiation pattern. This correction is made possible by altering the beam weights of the remaining elements in the array. The advantage of this method is that the necessity of replacement of the faulty elements is eliminated. Simulation is done on the linear antenna array constructed of individual isotropic elements separated by identical inter-element spacing and the results obtained from the simulation depict the effectiveness of the proposed method. This method can also be extended to other array geometries.

This paper presents a novel technique to design a time-modulator that can be used as a discrete component to generate different patterns in time-modulated antenna array (TMAA) system. The technique quantizes the modulation period into a number of levels with equal step size between the consecutive levels and rounds off the on-time of the antenna elements to nearest quantization levels to obtain quantized on-times (QOTs) of the corresponding elements. In order to realize the time modulation technique, a representative block diagram of a time-modulator defined as “quantized time-modulator” is presented. It is shown that the effect of quantization errors on the synthesized patterns can be diminished by selecting a suitable number of quantization states by using artificial bee colony (ABC) as a global optimizer. Accordingly, two previously reported examples based on simulated annealing (SA) algorithm, optimized by using continuous search space of on-time are considered, and ABC optimized results using discret...

A binary firefly algorithm inspired by the behavior of artificial fireflies has been applied to thin a planar concentric ring array antenna to minimize the peak sidelobe level in a number of predefined i¾?-planes. Two different cases comprising two different design specifications for the optimized thinned array have been considered. In the first case, the array is thinned to reduce the peak sidelobe level while keeping the first null beamwidth and half power beamwidth of the array below their specified values in all the defined i¾?-planes. In the second case, peak sidelobe level of the array is reduced regardless of any pre-specified values of first null beamwidth and half power beamwidth. To illustrate the superiority of binary firefly algorithm, the thinning of the array for the two different cases has been compared with binary particle swarm optimization algorithm. Results clearly show the superiority of binary firefly algorithm in finding optimum solutions for both cases. Copyright © 2014 John Wiley & Sons, Ltd.

The influence of evolutionary algorithms enhanced its scope of getting its existence in almost every complex optimization problems. In this paper, cuckoo search algorithm, an algorithm based on the brood parasite behavior along with Levy weights has been proposed for the radiation pattern correction of a linear array of isotropic antennas with uniform spacing when failed with more than one antenna element. Even though deterioration produced by the failure of antenna elements results in various undesirable effects, consideration in this paper is given to the correction of side lobe level and null placement at two places. Various articles in the past have already shown that the idea to correct the radiation pattern is to alter the amplitude weights of the remaining unfailed elements, instead of replacing the faulty elements. This approach is made use of modifying the current excitations of unfailed elements using cuckoo search algorithm such that the resulting radiation pattern is similar to the unfailed original pattern in terms of side lobe level and null placement at two places. Examples shown in this paper demonstrate the effectiveness of this algorithm in achieving the desired objectives.

In time-modulated antenna arrays (TMAAs), as a result of periodical switch-on and switch- ofi of the antenna elements, including operating frequency (termed as center frequency) signal, sideband signals are appeared at either sides of the center frequency in integer multiples of the modulation frequency. In this paper, it is shown that without using phase shifters, just by suitably controlling the on-time instants (OTIs) and on-time durations (OTDs) of a time-modulated linear antenna array (TMLAA) elements, simultaneously, along with a pencil beam pattern at the operating frequency, a shaped beam pattern can be obtained at both the flrst positive and negative harmonics of the time- modulation frequency. The important advantage of such a technique is that realization of multi-beam pattern in conventional antenna array (CAA) system generally requires complex feed network, whereas by using simple radio frequency (RF) switching circuit in the feed network of TMLAA, by virtue of the properties of harmonic radiations, synthesis of a shaped pattern at either (positive or negative) harmonic results in generating the same pattern in its opposite harmonic, and the synthesized patterns at difierent harmonics can be simultaneously used as independent communication channels. By employing a difierential evolution (DE) based optimization method, numerical results for a 16-element TMLAA with uniform excitation show that in conjunction with a pencil beam pattern at the center frequency, a ∞at-top or a cosec square pattern at flrst positive and negative sidebands of side-lobe levels (SLL) i20dB can be synthesized by suppressing the higher sideband level (SBL) to below i10dB.

A method based on Fast Fourier Transform for detecting the position and the level of fault in antenna arrays, from the degraded far field pattern samples is proposed. The paper considers Chebyshev array of specified side lobe level as a desired antenna pattern and authors develop strategy to test on it. The paper also deals with several different combinations of failures and successfully detects using the proposed technique, which takes lesser computational time than other optimization algorithms. In this paper, the method is developed and applied to a linear array of isotropic antennas; however, the principle can easily be extended to other type of arrays. 