Search

In order to improve the efficiency of antenna optimization design, a surrogate model is often used to replace the full-wave electromagnetic simulation software. Broad learning system (BLS) provides an alternative method for deep structure, aiming to overcome the drawback of excessive time-consuming training process, however, usually not with satisfactory accuracy. In order to further improve the performance of the model, master-apprentice (MA) behavior is proposed in this paper, using the current BLS training results as the priori knowledge, which are taken as fixed features to the next BLS hidden layer for further training. Each MA behavior forms a double BLS structure, which is composed of two parts, the models trained before and after are called master BLS (MBLS) and apprentice BLS (ABLS) respectively. These two subsystems together constitute a master-apprentice BLS (MABLS). Two antenna examples, rectangular microstrip antenna (RMSA) and WLAN dual-band monopole antenna (DBMA), and 10 UCI regression datasets are employed to demonstrate the effectiveness of the proposed model.

For antenna modeling and optimization design, high-fidelity full-wave electromagnetic simulation software can generally be used to obtain training samples. However, this process takes a long time. To solve this problem, a two-stage Gaussian process (GP) considering electromagnetic sensitivity information is proposed. Based on the coarse grid and fine grid of electromagnetic simulation software, the first stage learns the mapping relationships of antenna performance and sensitivity of each parameter, simultaneously. In the second stage, the accurate surrogates are established based on the mapping relationships above obtained. Because the method in this study takes sensitivity information into consideration during the modeling process, it can better reflect the mapping relationship between antenna input and output, which can develop a more accurate surrogate model. The proposed two-stage GP surrogate model considering sensitivity information can significantly reduce the demand of high-fidelity training samples for modeling, and greatly save the simulation time of calling electromagnetic simulation software. Therefore, this method is more suitable for the problems of insufficient electromagnetic simulation samples. The proposed approach is evaluated by modeling and optimization of an inverted F antenna and an ultra-wideband planar monopole antenna. The results show that the method has high modeling accuracy with limited training samples given by high-fidelity electromagnetic simulation, which further verify its effectiveness and efficiency.

In order to overcome the extremely time-consuming drawback of deep learning (DL), broad learning system (BLS) was proposed as an alternative method. This model is simple, fast, and easy to update. To ensure the fitting and generalization ability of BLS, the hidden layer neurons are often set too many, in fact, a lot of neurons are not needed. Greedy BLS (GBLS) is proposed in this paper to deal with the redundancy of the hidden layer in BLS from another perspective. Different from BLS, the structure of GBLS can be seen as a combination of unsupervised multi-layer feature representation and supervised classification or regression. It trains with a greedy learning scheme, performs principal component analysis (PCA) on the previous hidden layer to form a set of compressed nodes, which are transformed into enhancement nodes and then activated by nonlinear functions. The new hidden layer is composed of all newly generated compressed nodes and enhancement nodes, and so on. The last hidden layer of the network contains the higher-order and abstract essential features of the original data, which is connected to the output layer. Each time a new layer is added to the model, and there is no need to retrain from the beginning, only the previous layer is trained. Experimental results demonstrate that the proposed GBLS model outperforms BLS both in classification and regression.

Gaussian process (GP) is a rapidly developing supervised machine learning (ML) method in recent years, which has been widely used in the establishment of surrogate models in the field of electromagnetics. However, it has the problems of large sample demand, high computational complexity and low accuracy when processing high dimensional data. To solve this problem, a manifold Gaussian process (MGP) ML method based on differential evolution (DE) algorithm is proposed in this study. For the proposed method, the DE algorithm is used to get dimension reduction parameters, and the method can work very well with the optimized parameters. Compared with the traditional GP model, the dimensionality reduction method based on Isomap is adopted to simplify the mapping relationship between data pairs. Therefore, the model is more suitable for the problem of insufficient samples and high data dimension. In this study, the proposed DE-based MGP (DE-MGP) is applied to the extraction of coupling coefficients of the fourth-order and sixth-order coupling filters, in which the test error of the fourth-order coupling filter surrogate model can be reduced to 0.84%, and the test error of the sixth-order coupling filter is expected to be reduced to 1.53%, which proves that the proposed method is very effective.

In order to design the antenna that meets the requirements conveniently and reduce the tedious manual operation in the process of antenna optimization, a new antenna design method is proposed in this paper, which calls the Gaussian process surrogate model of antenna through differential evolution algorithm. According to the proposed method, when designing the antenna, we not only can optimize the physical parameters by intelligent optimization algorithm, but also avoid the huge time consumption caused by using electromagnetic simulation software. In this study, we apply the proposed method to the design experiment of WlAN dual band monopole antenna, and finally get a set of physical parameters, by these parameters, the antenna model response meets the design requirements.

In this paper, a symmetric stepped-impedance coupled line (SSICL) is proposed to realize dual-band quarter-wave coupled line (DQCL), giving the same equivalent even and odd characteristic impedances at two different frequencies. The DQCL can be directly used to design dual-band parallel-line coupler (DBPC) with arbitrary coupling coefficient and large frequency ratio, resulting in even shorter total length when compared with traditional quarter-wave parallel-line coupler working at lower frequency $f_{1}$ . In addition, broad DBPC (BDPC) is constructed by cascading multiple sections of SSICLs, as the SSICL is quite similar with quarter-wave coupled line in circuit. For demonstrating the design methods, in the first step, two edge-coupled 20-dB DBPCs and one broadside coupled 6-dB DBPC, working at 1/5.2, 1/7.5, and 1/5.2 GHz, respectively, are designed, fabricated and measured, exhibiting good features like large frequency ratio, high isolations, and compact dimensions. Then, for broadening bandwidths, three-section BDPC working at 1 GHz/5.2 GHz is designed and measured, resulting in working bandwidths larger than 1.2 GHz. These good features have proved the feasibility of the proposed circuits and the corresponding design methods.

Aiming at the problem of Gaussian process machine learning to solve the problem of low efficiency caused by too many samples in the process of antenna modeling, a Gaussian process modeling method based on prior knowledge is proposed. The input sample of the model is the size parameter of the antenna, and the output sample is the difference of the electromagnetic simulation calculation result and the prior knowledge, thus establishing the Gaussian process model. This modeling method is used to model the resonant frequency of the coplanar waveguide feed dual-frequency microstrip antenna, and the error result is ideal. It is shown that the Gaussian process model based on prior knowledge can replace the electromagnetic simulation software, improve the modeling speed, and prove the feasibility of the method.

Electromagnetic simulation has become an important tool in modern antenna structure design. However, accurate simulations of real antenna models are expensive, so designing automation by using EM solvers in optimized antennas may be prohibitive because of its high computational cost. In this article, a gaussian process model as the fitness function of the particle swarm algorithm instead of electromagnetic simulation software to optimize the microwave dual-band monopole antenna has been demonstrated.

In order to improve the efficiency of optimal design of microstrip antennas (MSAs), this paper proposes two methods. One is the Gaussian Process Particle Swarm Optimization (GP-PSO) algorithm, in which Gaussian Process (GP) model is used to replace the full-wave electromagnetic simulation software, and Particle Swarm Optimization (PSO) algorithm is used for optimization. This method is applied to optimize a GPS Beidou dual-mode MSA. The results show that the antenna designed by this method not only meets the design specifications but also greatly reduces the optimization time. In order to increase the computing efficiency of the proposed GP-PSO algorithm further, a second method named fitness-estimation-based PSO (FE-PSO) algorithm is proposed. According to the explicit evolution formula of PSO algorithm, the prediction formula of the fitness value of particles is constructed, and the fitness value can be predicted by the formula from the third generation of the proposed FE-PSO algorithm. The formula gives the prediction result instead of the GP model in the GP-PSO algorithm, aiming to reduce the modeling time. A Yagi MSA is optimized by this method. The results indicate that in the case of ensuring the design accuracy it can achieve the purpose of rapid optimization.

Novel lumped dual-frequency impedance transformers (DFITs) for arbitrary frequency dependent complex loads (FDCLs) are synthesized using the concepts of capacitive and inductive inverters. Such a dual-frequency impedance transformer is constructed with one Pi-shaped lumped inverter and two shunt susceptance blocks at two sides, and get simplified after susceptance merging. The shunt susceptance blocks have specified susceptance values at two frequencies and are realized with lumped circuits of certain topologies. Design formulas of the DFITs are derived and are validated by two examples. Compared with a former technique, this method uses less lumped elements and provides one design freedom, so largest matching bandwidths can be achieved by comparing a series of possible designs.

We present a novel binary version of multiobjective wind driven optimization (WDO) for emitted beamforming of opportunistic array radar, which is assumed as a multiobjective optimization problem. Firstly, the emitted signal model and objective functions of optimization are presented. Then the algorithm proposes a new definition of the position vector of air parcel, and brings a good discretization interpretation of continuous WDO. For multiobjective optimization, the grey relational grade (GRG) is then used to measure the similarity between the best two solutions for these two objectives. The best pressure locations with the maximum GRG will be recorded as the best two candidate solutions to the problem, and a final optimization result will be selected according to the importance of the two objectives. Finally, the proposed improved WDO has been applied for the optimal design of beamforming of the opportunistic antenna array, which needs a trade-off between the 3 dB main beam width and sidelobe level. The simulation results show that the proposed method outperforms conventional particle swarm optimization (PSO) in the optimal beamforming by achieving more reduction in the sidelobe level and saving more runtime.

When microwave devices are designed by knowledge-based neural network (KBNN), the empirical formula is always used as priori knowledge. However, it is difficult to derive the corresponding formulas for the most electromagnetic problems, especially for complex electromagnetic problems, the formula derivation is almost impossible. In this article, they combine neural network with simulation software and use results of Agilent ADS as priori knowledge and HFSS as teaching signal to train the neural network by particle swarm optimization (PSO), which solves the difficulty in obtaining priori knowledge and effectively reduces the complexity of the neural network structure. Based on the KBNN, the microwave filters are designed. The results of optimization satisfy the required specifications which show the effectiveness and superiority of the method.

This paper proposes a novel image magnification method based on bilinear interpolation, wavelet, and partial differential equation (PDE) techniques. The image which is interpolated linearly is decomposed by wavelet into a low frequency component image and three high frequency component images, and then the three high frequency component images and the original image regarded as low-frequency component will be used for image magnification by invert wavelet transform. Finally, a PDE involving gray fidelity constraint item called improvement-self-snake mode is presented in post-processing of the magnified image. The experimental results show that the proposed linear interpolation-wavelet-PDE approach is indeed efficient and effective in image magnification. In addition, we also compare the signal-to-noise ratio (SNR) of the linear interpolation-wavelet-PDE magnification method with methods of linear interpolation, linear interpolation-wavelet, and wavelet-PDE. The simulating results show that the linear interpolation-wavelet-PDE method indeed outperforms the three kinds of image magnification approaches mentioned above.

This paper proposes a method to extract speech features for content integrity verification of G. 729 coded speech, which could minimize the computation cost and reduce the sizes of the features. Experiments results show that the method is robust to common operations in speech transmission such as recompression, volume changing, re-sampling and transcoding, and also does well in tamper detection for its ability to distinguish content preserving operations from malicious modifications.