Research Progress on Rapid Optimization Design Methods of Metamaterials Based on Intelligent Algorithms
-
摘要: 目前,超材料研究不断向工程化应用推进,在物理机理与效应、设计理论与方法、加工制备与测试等方面取得了突飞猛进的发展。但是,传统的超材料设计主要依赖人工设计和优化,面对大规模的工程化应用设计时,无法实现数量庞大的超材料结构单元的快速整体设计。近几年,涵盖传统启发式算法和神经网络算法的智能算法在超材料设计中所占的比重逐步上升,基于智能算法设计超材料能够打破传统设计方法在不同基材体系、不同频段以及不同性能指标下设计的局限性,展现出快速设计和架构创新的独特优势。该文综述了包括遗传算法、Hopfield网络算法和深度学习在内的几种典型智能算法在超材料设计中的应用,包括正向设计方法和逆向设计方法。基于智能算法能够实现不同性能指标的频率选择表面、多机理复合吸波超材料、平板聚焦超表面以及异常反射超表面的快速设计,为推动超材料技术的工程化应用提供必要设计手段支撑。Abstract: At present, research on metamaterials is continuously advancing to engineering applications, and great progress is being achieved in the areas of physical mechanisms and effects, design theory and methods, and fabrication and measurement. However, traditional metamaterials design mainly relies on artificial design and optimization. In the face of large-scale engineering applications, it is impossible to realize the rapid overall design of a large number of metamaterial structural units. In recent years, the proportion of intelligent algorithms covering traditional heuristic algorithms and neural network algorithms in metamaterials design has increased gradually. Metamaterials design based on intelligent algorithms can surpass the limitation of traditional methods in different substrate systems, frequency variation, and different performance indicators, offering the unique advantages of rapid design and architectural innovation. This paper summarizes the application of several typical intelligent algorithms, including the genetic algorithm, Hopfield network algorithm, and deep learning algorithm in metamaterials design, which include forward designs and an inverse design. The use of intelligent algorithms can achieve the rapid design of frequency selective surfaces under different performance indexes, multi-mechanisms composite absorber metamaterials, flat focusing, and abnormal reflection metasurfaces, providing the necessary support for design methods while promoting the engineering applications of metamaterials.
-
1. Introduction
High resolution radar imaging has been widely used in target scattering diagnostics and recognition. As we all know, high resolution in range dimension is derived from the bandwidth of the transmitting signal and in the cross range dimension from synthetic aperture of multiple spatial positions. Under the fixed bandwidth and the synthetic aperture, traditional Matched Filter (MF) based methods for radar imaging suffer from low resolution and high sidelobes limited by the synthetic aperture[1].
In order to improve the resolution and suppress the sidelobes, many high resolution methods have been applied to radar imaging. For example, the recently introduced theory of Compressed Sensing (CS) provides an idea to improve the resolution and reduce the amounts of measurement data under the constraint of sparsely distributed target prior, which has been widely explored for applications of radar imaging[2–4]. However, conventional CS methods are confronted with a range of problems in practical scenarios, such as complexity in calculation, high Signal-to-Noise Ratio (SNR) requirement, model mismatch caused by off grid problem[5], phase mismatch[6], frequency error[7] and position error[8]. To avoid the off grid problem of CS, modern spectral estimation methods like MUltiple SIgnal Classification (MUSIC), matrix pencil and Estimation of Signal Parameters via Rotational Invariance Techniques (ESPRIT) have been used in radar imaging for resolution improvement[9]. However, most those methods suffer from performance degradation when there is little prior knowledge of the exact numbers of the scatters or under low SNR condition. Recently, the atomic norm minimization algorithm[10] based on continuous compressed sensing is introduced to enhance the SNR of the received echo and using Vandermonde decomposition to eliminate the grid mismatch. Nevertheless, this method can only be tailored to a specific model and brings huge computational cost.
Consideration the aforementioned fact while combining the sparsity low rank matrix recovery technology and deconvolution algorithm, we introduce a high resolution radar imaging method based on the MF result. Firstly we establish the convolution model of target’s backscatter coefficients and the Point Spread Function (PSF), and then we want to use the deconvolution method like Wiener filter to improve the radar imaging resolution. However, the performance improvements of those methods depend on high SNR, and their super resolution performance is visibly affected by the low pass character of the PSF[11]. Although the MF result has enhanced the SNR, we can further improve the echo SNR by the sparsity and low rank matrix recovery. Low rank matrix recovery has been applied in many signal processing applications to estimate a low rank matrix from its noisy observation[12, 13]. Combinng the sparsity of the echo matrix, we modify the low rank matrix recovery and introduce it to radar echo denoising, which can improve the performance of the two-Dimensional (2D) deconvolution. Finally, some experimental results are conducted to verify the effectiveness of the proposed method.
Notation: (·)T, (·)H and (·)* denote the transpose, the conjugate transpose and the conjugate operation, respectively.
⟨⋅⟩ ,﹡, and ☉ indicate the inner product, the convolution and the Hadamard product.‖⋅‖F ,‖⋅‖1 ,‖⋅‖∗ are the Frobenius norm, sum of the absolute values and the nuclear norm.2. Model Establishment
Considering a typical arrangement for radar imaging in which an object with scattering reflectivity
σxy rotated by a scan angleθm (as shown in Fig. 1), we defined the positions of the transmitting and receiving antenna shown in a Cartesian coordinate as (W/2, –R, H) and (–W/2, –R,H), where W, R, H represent the antenna spacing, distance from antenna to XZ plane and XY plane, respectively.Transmitting a stepped-frequency signal with frequency
fn and under Born approximation, the received scattered echo with Gauss noise Wmn is given by:Ymn=∬Sσxye−j2πfnR(x,y;θm)/cdxdy+Wmn (1) In this equation:
fn=f0+nΔf,n=0,1···,N−1 ,f0 andΔf represent the start frequency and frequency step,θm=mΔθ,m=0,1,···,M ,Δθ represents the rotating angle step, respectively. The rangeR(x,y;θm) from the transmitting antenna to echo scattering center and to the receiving antenna can be calculated as Eq. (2):R(x,y;θm)=√(xcosθm−ysinθm+W2)2+(xsinθm+ycosθm+R)2+H2+√(xcosθm−ysinθm−W2)2+(xsinθm+ycosθm+R)2+H2 (2) In far-field and small rotation angle case,
R(x,y;θm) can be approximated by first order Taylor-series expansion as:R(x,y;θm)≈2(R0+(x+mΔθy)R/R0) (3) where
R0=√R2+(W/2)2+H2 .Then the received echo can be written as follow under some approximated conditions:
˜Ymn=∬S˜σxye−j4πRΔθλR0mxe−j4πRΔfR0cnydxdy+Wmn (4) where
˜σxy=σxye−j4πRy/λR0 ,˜Ymn=Ymnej4πfnR0/c ,λ≈c/f0 .After discrete imaging region with P×Q grids, the received echo in Eq. (4) can be described as the following 2D linear signal model:
˜Y=Ax˜ΣATy+W (5) where
˜Y=[ ˜Ymn]M×N is the echo matrix,˜Σ=[˜σpq]P×Q is the observation matrix:Ax=[e−j4πmxpRΔθ/λR0]M×P ,Ay=[e−j4πnyqRΔf/R0c]N×Q .Considering the targets present sparse point scattering characteristic under high frequency scattering in most practical application scenarios, we present our method to improve the resolution of radar imaging under sparse target constraint using 2D deconvolution algorithm with low rank sparsity echo matrix denoising.
3. 2D Deconvolution with Echo Denoising
As we all know, the MF algorithm which is based on the maximum signal to noise ratio is the most stable and commonly used radar imaging method. However, due to limitation of the synthetic aperture and bandwidth, the standard MF method suffers from relatively low resolution and high sidelobes, especially under the requirements of high resolution. The received echo after MF from Eq. (5) can be obtained by:
YMF=AHx ˜YA∗y (6) From the result of Eq. (6), the echo of the surface target after MF can be described as the sum of all the wave scattered at the points on the surface grid, i.e.,
YMF(x,y)=∑x′∑y′˜σ(x′,y′)Psf(x−x′,y−y′) (7) where we define the PSF as:
Psf(x−x′,y−y′)=⟨ax(x),ax(x′)⟩⟨ay(y),ay(y′)⟩ (8) here,
ax(x) anday(y) represent the column of matrix Ax and Ay at the grid (x, y).We can find that Eq. (6) can be seen as the 2D convolution of the PSF and target backscatter coefficients:
YMF(x,y)=˜σ(x,y)∗Psf(x,y)+WMF(x,y) (9) Inspired by this, we can recovery the backscatter coefficients using deconvolution algorithm to improve the imaging quality. Firstly, we should analyze the characteristic of the PSF and its influence on the deconvolution result.
3.1 Analysis of PSF
The PSF can be evaluated as:
Psf(x,y)≈e−j2π[(M−1)RΔθλR0x+(N−1)RΔfR0cy]⋅sinc(2MRΔθλR0x)sinc(2NRΔfR0cy) (10)
We can calculate the 2D mainlobe width which represents the radar imaging resolution as follows:
ρx=λR02MRΔθ,ρy=R0c2NRΔf (11) Eq. (9) indicates that the MF result can be seen as the convolution result of backscatter coefficients and
Psf(x,y) . The PSF is characterized by synthetic aperture and bandwidth, which has strong low pass characteristic with low resolution and high sidelobes as shown in Eq. (10) and Eq. (11). For an isolated target scatter, imaging result after the MF output will be proportional to the PSF, therefore the resolution of MF result is limited and accompanied by low resolution and high sidelobes. Inspired by above, we can restore the high resolution backscatter coefficients information by deconvolution to remove the effect of low pass characteristic of PSF.3.2 2D deconvolution
As we have get 2D convolution form as Eq. (9), here we consider to use the direct deconvolution algorithm to recovery target backscatter coefficients. Firstly, we transform Eq. (9) into the spatial frequency domain using 2D Fourier transform as:
Yω=Σω⊙Hωω+Wω (12) where,
Yω=F{YMF},Σω=F{˜Σ},Hω=F{Psf}, Wω=F{WMF} .Theoretically, the target scattering information could be restored by deconvolution as:
Σω=Yω/Hω (13) However,
1/Hω will be very large in practice at the outside of the mainlobe of PSF since the low pass characteristic of the PSF, which results in tremendous amplification of noise and obtains valueless results. So the deconvolution processing becomes an ill-posed inverse problem.In order to alleviate the ill-posed problem, we use Winner filtering algorithm and sparse low rank matrix recovery to improve the quality of imaging result.
The result after Winner filter algorithm can be written as[14]:
˜Σω=Yω⊙Hω∗‖Hω‖2+ΨWW(ω)/ΨΣΣ(ω) (14) where
ΨWW(ω) andΨΣΣ(ω) is the power spectral density ofW andΣ . Eq. (14) will approach Eq. (13) when the SNR is relatively high. What’s more, Eq. (14) will attenuate the high frequencies noise to alleviate the ill-posed problem under low SNR. In experimental data processing, theΨWW(ω)/ΨΣΣ(ω) is generally set according to the experience value. We can get the scattering reflectivity˜Σ by 2D Inverse Fourier transform according to Eq. (14).3.3 Echo denoising by sparse low rank matrix recovery
We can prove that the echo matrix after MF is sparse and low rank in Appendix A and by using this characteristic, the echo SNR can be improved. Consider the problem of estimating a sparse low rank matrix X from its noisy observation Y:
Y=X+W (15) Define the sparse low rank matrix recovery problem as:
min (16) where
\gamma is the regularization parameter used to balance the relative contribution between nuclear norm and the 1-norm, which can control the denoising performance. In general, the denoising threshold of\gamma can be set as the 5%~10% of the maximum singular value of Y.By applying Augmented Lagrangian Method (ALM), we can get the optimization problem:
\begin{align} F\left( {{{\text{X}}},{{\text{D}}},{{{\text{Y}}}_{\!1}},{{{\text{Y}}}_{\!2}},\mu } \right) \!=\! & \gamma {\left\| {{\text{X}}} \right\|_*} + \left\langle {{{{\text{Y}}}_{\!1}},{{\text{Y}}} - {{\text{X}}}} \right\rangle \\ &+\! \frac{\mu }{2}\left\| {{{\text{Y}}} \!-\! {{\text{X}}}} \right\|_F^2 \!+\! \left( {1 \!-\!\! \gamma } \right){\left\| {{\text{D}}} \right\|_1} \\ &+\! \left\langle {{{{\text{Y}}}_2},{{\text{D}}} - {{\text{X}}}} \right\rangle \!\!+\!\! \frac{\mu }{2}\left\| {{{\text{D}}} \!\!-\!\! {{\text{X}}}} \right\|_F^2 \end{align} (17)
And the update rules for solving this problem are as follows:
{{{\text{X}}}^{\left( {k + 1} \right)}} = \mathcal{S}\left( {\frac{{{{\text{Y}}} \!\!+\!\! {{{\text{D}}}^{\left( k \right)}}}}{2} \!+\! \frac{{{{\text{Y}}}_1^{\left( k \right)} \!+\! {{\text{Y}}}_2^{\left( k \right)}}}{{2{\mu ^{\left( k \right)}}}},\frac{\gamma }{{2{\mu ^{\left( k \right)}}}}} \right) (18) {{{\text{D}}}^{\left( {k + 1} \right)}} = {\rm{soft}}\left( {\frac{1}{{{\mu ^{\left( k \right)}}}}{{\text{Y}}}_2^{\left( k \right)} - {{{\text{X}}}^{\left( {k + 1} \right)}},\frac{{1 - \gamma }}{{{\mu ^{\left( k \right)}}}}} \right)\quad (19) \!\!\!\!\!\!\!\!\!\!\left. \begin{align} {{\text{Y}}}_1^{\left( {k + 1} \right)} =& {{\text{Y}}}_1^{\left( k \right)} + {\mu ^{\left( k \right)}}\left( {{{\text{Y}}} - {{{\text{X}}}^{\left( {k + 1} \right)}}} \right) \\ {{\text{Y}}}_2^{\left( {k + 1} \right)} =& {{\text{Y}}}_2^{\left( k \right)} + {\mu ^{\left( k \right)}}\left( {{{{\text{D}}}^{\left( {k + 1} \right)}} - {{{\text{X}}}^{\left( {k + 1} \right)}}} \right) \quad\; \\ {\mu ^{\left( {k + 1} \right)}} =& \beta {\mu ^{\left( k \right)}},\,\,\, \beta > 1 \end{align}\!\!\!\!\!\!\right\} (20) where,
\mathcal{S}\left( { \cdot , \cdot } \right) is the singular value thresholding function defined as:\mathcal{S}\left( {{{\text{X}}},\gamma } \right){\rm{ = }}{{\text{U}}}{\rm{soft}}\left( {{\text{Σ}},\gamma } \right){{{\text{V}}}^{\rm T}} (21) where,
{{\text{X}}} = {{\text{U}}}{\text{Σ}}{{{\text{V}}}^{\rm T}\,} is the Singular Value Decomposition (SVD) of{{\text{X}}} ,{\rm{soft}}\left( \cdot \right) is the soft thresholding function defined as:{\rm{soft}}\left( {x,\gamma } \right) = {\rm{sign}}\left( x \right) \cdot \max \left\{ {\left| x \right| - \gamma ,0} \right\} (22) See Appendix B for the detailed derivation of Eq. (18) and Eq. (19).
The flowchart of the proposed method is shown in Fig. 2 by combining the sparse low rank matrix recovery with the 2D deconvolution.
4. Simulation
The parameters in the simulation are given in Tab. 1. In this experiment, we set four-point targets, the imaging results are shown in Fig. 3.
Table 1. Simulation parametersParameter Value Parameter Value {{M}} 256 R 1 m {{N}} 500 H 0.7 m \Delta f 10 MHz W 0.04 m \Delta \theta 0.009° SNR –15 dB As shown in Fig. 3(a), due to the limitation of synthetic aperture and bandwidth, the MF method suffers from relatively low resolution and high sidelobes which make it difficult to distinguish between four-point targets even there is no noise. Fig. 3(b)–Fig. 3(d) show the imaging results reconstructed by MF and proposed method including the intermediate denoising results when SNR = –15 dB. It can be clearly seen that the effect of denoising compared Fig. 3(c) with Fig. 3(a) and Fig. 3(b), the echo SNR is further improved by the sparsity and low rank matrix recovery during the proposed intermediate denoising procedure. The final imaging result is shown in Fig. 3(d), from which we can see that the proposed method has a better reconstruction precision with higher resolution imaging of four distinguishable point targets.
5. Experiment Results
The experimental scene is shown in Fig. 4(a), which is the same with the model in Fig. 1. The radar system consists of a pair of horn antennas, a turntable whose rotation angle can be precisely controlled by the computer, and an Agilent VNA N5224A which is used for transmitting and receiving the stepped-frequency signal with bandwidth of 10 GHz from 28 GHz to 38 GHz and number of frequencies N equals to 256 (Frequency interval
\Delta f is 40 MHz). Two kind of targets including three 5-mm-diameter mental spheres and a pair of scissors placed on a rotatory platform are used here as shown in Fig. 4(b).As we know, image entropy can be considered as a metric for measuring the smoothness of the probability density function of image intensities[15]. The imaging entropy is defined as:
E\left( I \right) = - \sum\limits_{p = 1}^P {\sum\limits_{q = 1}^Q {\left| {\frac{{{I^2}\left( {p,q} \right)}}{{s\left( I \right)}}} \right| \ln \left| {\frac{{{I^2}\left( {p,q} \right)}}{{s\left( I \right)}}} \right|} } (23) where
s\left( I \right) = \displaystyle\sum\nolimits_{p = 1}^P {\sum\nolimits_{q = 1}^Q {{{\left| {I\left( {p,q} \right)} \right|}^2}} } .5.1 Three mental spheres
In this experiment, we set
R \!=\! 1\;{\rm m},\ H \!=\! 0.7\;{\rm m}, W = 0.04\;{\rm m} and the total rotating angle is{5^ \circ } with an angle interval\Delta \theta = {0.01^ \circ } (M = 500 ).Fig. 5 shows the results of the MF and our proposed method for the mental spheres. The one-dimensional
x andy domain cross-section of the target with red-dashed circle shown in Fig. 5 are presented in Fig. 6, in which the red-dashed line and blue line represent the result of MF and proposed method. Clearly, the reconstruction result of proposed method has a narrower main-lobe and lower side-lobe than MF and the sharpening ratio almost reach 5.8 and 3.0 inx andy domain, respectively.5.2 Scissors
The parameters for this experiment are set as follows,
R = 0.876\;{\rm m} ,W = 0.04\;{\rm m} . The total rotating angle is{360^ \circ } withM equals to{720^ \circ } . Taking into account the scintillation characteristics of the target under large rotating angle, we divide the rotating angle into 72 segments and each of the part is with rotating angle from{0^ \circ } to{5^ \circ } . The proposed method is used to process the data for each segment and the image fusion method is used to merge the results of all segments.Fig. 7 shows the imaging results of the scissors reconstructed by MF and proposed method. It can be seen from the results that the proposed method has a high reconstruction precision with a shaper shape of scissors.
The entropies of the imaging results by MF and our proposed method are given in Tab. 2 to quantitatively assess the performance. The proposed method has a low entropy which means the proposed method can improve the resolution and verifies its superiority.
Table 2. Entropies of imaging resultsTarget MF Our proposed method Mental spheres 8.7282 4.8429 Scissors 8.9433 7.0454 6. Conclusion
We introduce a robust deconvolution method with enhancing SNR technology to realize high resolution radar imaging. Compared to other high resolution methods, our proposed method is simple and robust. Although the signal model and experiments are performed for turntable radar situation with SF waveform, the method can be directly generalized to other practical radar systems based on other types of signals.
Appendix A Proof of the sparsity and low rank characteristic
To prove the echo matrix after MF is sparse and low rank, the following lemma is needed.
Lemma 1[16]: For matrix A and B, the ranks of the product of A and B satisfy the inequality below:
{\rm{rank}}\left( {{{\text{AB}}}} \right) \le \min \left\{ {{\rm{rank}}\left( {{\text{A}}} \right),{\rm{rank}}\left( {{\text{B}}} \right)} \right\} (A-1) From Eq. (5) and Eq. (6), we can see that the echo matrix after MF can be written as:
{{{\text{Y}}}_{\!\!{\rm {MF}}}} \ = {{\text{A}}}_x^{\rm H}{{{\text{A}}}_x}\tilde{{\text{Σ}} }{{\text{A}}}_y^{\rm T}{{\text{A}}}_y^* (A-2) We have supposed that the target has sparse distribution, so the target backscatter coefficients matrix
\tilde{{\text{Σ}} } is sparse and low rank. Thus, matrix{{{\text{Y}}}_{{\!\rm{MF}}}} is also low rank according to lemma 1. The sparsity of matrix{{{\text{Y}}}_{{\rm{\!\!MF}}}} can be proved by Eq. (9) obviously.Appendix B Derivation of Eq. (18) and Eq. (19)
For Eq. (18), the optimization problem can be described as Eq. (B-1), and it has a closed-form solution just as Eq. (18) according to Ref. [13].
\begin{align} {{{\text{X}}}^{\left( {k + 1} \right)}}\!\!\!\!\!\!\!\!\!\!\! & \\ & =\!\! \arg \mathop {\min }\limits_{{\text{X}}} F\left( {{{\text{X}}},{{{\text{D}}}^{\left( k \right)}},{{\text{Y}}}_1^{\left( k \right)},{{\text{Y}}}_2^{\left( k \right)},{\mu ^{\left( k \right)}}} \right) \\ & =\!\! \arg \mathop {\min }\limits_{{\text{X}}} \frac{1}{2}\left\| {{\text{X}}} \!- \! \frac{1}{2}\left( {{{\text{Y}}} \!+\! {{{\text{D}}}^{\left( k \right)}} \!+\! \frac{{\text{Y}}_1^{\left( k \right)}\!+\!\! {{\text{Y}}}_2^{\left( k \right)}}{{{\mu ^{\left( k \right)}}}}} \right) \right\|_F^2 \\ &\quad + \frac{\gamma }{{2{\mu ^{\left( k \right)}}}}{\left\| {{\text{X}}} \right\|_*} \end{align} (B-1) For Eq. (19), it is the same with Eq. (18), which can written as
\begin{align} {{{\text{D}}}^{\left( {k + 1} \right)}} = & \arg \mathop {\min }\limits_{{\text{D}}} F\left( {{{{\text{X}}}^{\left( {k + 1} \right)}},{{\text{D}}},{{\text{Y}}}_1^{\left( k \right)},{{\text{Y}}}_2^{\left( k \right)},{\mu ^{\left( k \right)}}} \right) \\ = & \arg \mathop {\min }\limits_{{\text{D}}} \frac{1}{2}\left\| {{{\text{D}}} - \left( {\frac{{{{\text{Y}}}_2^{\left( k \right)}}}{{{\mu ^{\left( k \right)}}}} - {{{\text{X}}}^{\left( {k + 1} \right)}}} \right)} \right\|_F^2 \\ &+ \frac{{1 - \gamma }}{{{\mu ^{\left( k \right)}}}}{\left\| {{\text{D}}} \right\|_1} \end{align} (B-2) It also has a closed-form solution as Eq. (19) according to Ref. [17].
-
-
[1] PENDRY J B. A chiral route to negative refraction[J]. Science, 2004, 306(5700): 1353–1355. doi: 10.1126/science.1104467 [2] PENDRY J B, HOLDEN A J, STEWART W J, et al. Extremely low frequency plasmons in metallic mesostructures[J]. Physical Review Letters, 1996, 76(25): 4773–4776. doi: 10.1103/PhysRevLett.76.4773 [3] PENDRY J B, HOLDEN A J, ROBBINS D J, et al. Magnetism from conductors and enhanced nonlinear phenomena[J]. IEEE Transactions on Microwave Theory and Techniques, 1999, 47(11): 2075–2084. doi: 10.1109/22.798002 [4] SMITH D R, PADILLA W J, VIER D C, et al. Composite medium with simultaneously negative permeability and permittivity[J]. Physical Review Letters, 2000, 84(18): 4184–4187. doi: 10.1103/PhysRevLett.84.4184 [5] CUI Tiejun, LIU Shuo, and LI Lianlin. Information entropy of coding metasurface[J]. Light: Science & Applications, 2016, 5(11): e16172. [6] XIE Boyang, TANG Kun, CHENG Hua, et al. Coding acoustic metasurfaces[J]. Advanced Materials, 2017, 29(6): 1603507. doi: 10.1002/adma.201603507 [7] 史峰, 王辉, 郁磊, 等. MATLAB智能算法30个案例分析[M]. 北京: 北京航空航天大学出版社, 2011.SHI Feng, WANG Hui, YU Lei, et al. Analysis of 30 Cases of MATLAB Intelligent Algorithm[M]. Beijing: Beijing University of Aeronautics and Astronautics Press, 2011. [8] 武飞周, 薛源. 智能算法综述[J]. 工程地质计算机应用, 2005, (2): 9–15.WU Feizhou and XUE Yuan. Review of intelligent algorithms[J]. Engineering Geology Computer Application, 2005, (2): 9–15. [9] 胡涵, 李振宇. 多目标进化算法性能评价指标综述[J]. 软件导刊, 2019, 18(9): 1–4. doi: 10.11907/rjdk.191024HU Han and LI Zhenyu. A survey of performance indicators for multi-objective evolutionary algorithms[J]. Software Guide, 2019, 18(9): 1–4. doi: 10.11907/rjdk.191024 [10] 梅志伟. 多目标进化算法综述[J]. 软件导刊, 2017, 16(6): 204–207. doi: 10.11907/rjdk.171169MEI Zhiwei. Overview of multi objective evolutionary algorithm[J]. Software Guide, 2017, 16(6): 204–207. doi: 10.11907/rjdk.171169 [11] CAI Haoyuan, SUN Yi, WANG Xiaoping, et al. Design of an ultra-broadband near-perfect bilayer grating metamaterial absorber based on genetic algorithm[J]. Optics Express, 2020, 28(10): 15347–15359. doi: 10.1364/OE.393423 [12] HUANG Yixing, FAN Qunfu, CHEN Jin, et al. Optimization of flexible multilayered metastructure fabricated by dielectric-magnetic nano lossy composites with broadband microwave absorption[J]. Composites Science and Technology, 2020, 191: 108066. doi: 10.1016/j.compscitech.2020.108066 [13] QIU Tianshuo, SHI Xin, WANG Jiafu, et al. Deep learning: A rapid and efficient route to automatic metasurface design[J]. Advanced Science, 2019, 6(12): 1900128. doi: 10.1002/advs.201900128 [14] TITTL A, LEITIS A, LIU Mingkai, et al. Imaging-based molecular barcoding with pixelated dielectric metasurfaces[J]. Science, 2018, 360(6393): 1105–1109. doi: 10.1126/science.aas9768 [15] MA Wei, CHENG Feng, and LIU Yongmin. Deep-learning enabled on-demand design of chiral metamaterials[J]. ACS Nano, 2018, 12(6): 6326–6334. doi: 10.1021/acsnano.8b03569 [16] LIU Che, YU Wenming, MA Qian, et al. Intelligent coding metasurface holograms by physics-assisted unsupervised generative adversarial network[J]. Photonics Research, 2021, 9(4): B159–B167. doi: 10.1364/PRJ.416287 [17] PEURIFOY J, SHEN Yichen, YANG Yi, et al. Nanophotonic inverse design using artificial neural network[C]. Frontiers in Optics 2017, Washington, USA, 2017: FTh4A.4. doi: 10.1364/FIO.2017.FTh4A.4. [18] VAI M M, WU Shuichi, LI Bin, et al. Reverse modeling of microwave circuits with bidirectional neural network models[J]. IEEE Transactions on Microwave Theory and Techniques, 1998, 46(10): 1492–1494. doi: 10.1109/22.721152 [19] KABIR H, WANG Ying, YU Ming, et al. Neural network inverse modeling and applications to microwave filter design[J]. IEEE Transactions on Microwave Theory and Techniques, 2008, 56(4): 867–879. doi: 10.1109/TMTT.2008.919078 [20] SELLERI S, MANETTI S, and PELOSI G. Neural network applications in microwave device design[J]. International Journal of RF and Microwave Computer-Aided Engineering, 2002, 12(1): 90–97. doi: 10.1002/mmce.7001 [21] LIU Dianjing, TAN Yixuan, KHORAM E, et al. Training deep neural networks for the inverse design of nanophotonic structures[J]. ACS Photonics, 2018, 5(4): 1365–1369. doi: 10.1021/acsphotonics.7b01377 [22] PEURIFOY J, SHEN Yichen, JING Li, et al. Nanophotonic particle simulation and inverse design using artificial neural networks[J]. Science Advances, 2018, 4(6): eaar4206. doi: 10.1126/sciadv.aar4206 [23] 随赛. 新型人工电磁表面拓扑优化设计与应用研究[D]. [博士论文], 空军工程大学, 2019.SUI Sai. Research on topology optimization design and application of new artificial electromagnetic surface[D]. [Ph. D. dissertation], Air Force Engineering University, 2019. [24] LANDY N I, SAJUYIGBE S, MOCK J J, et al. Perfect metamaterial absorber[J]. Physical Review Letters, 2008, 100(20): 207402. doi: 10.1103/PhysRevLett.100.207402 [25] 鲍迪, 沈晓鹏, 崔铁军. 太赫兹人工电磁媒质研究进展[J]. 物理学报, 2015, 64(22): 228701. doi: 10.7498/aps.64.228701BAO Di, SHEN Xiaopeng, and CUI Tiejun. Progress of terahertz metamaterials[J]. Acta Physica Sinica, 2015, 64(22): 228701. doi: 10.7498/aps.64.228701 [26] SHEN Yang, ZHANG Jieqiu, PANG Yongqiang, et al. Transparent broadband metamaterial absorber enhanced by water-substrate incorporation[J]. Optics Express, 2018, 26(12): 15665–15674. doi: 10.1364/OE.26.015665 [27] PANG Yongqiang, SHEN Yang, LI Yongfeng, et al. Water-based metamaterial absorbers for optical transparency and broadband microwave absorption[J]. Journal of Applied Physics, 2018, 123(15): 155106. doi: 10.1063/1.5023778 [28] PANG Yongqiang, LI Yongfeng, WANG Jiafu, et al. Electromagnetic reflection reduction of carbon composite materials mediated by collaborative mechanisms[J]. Carbon, 2019, 147: 112–119. doi: 10.1016/j.carbon.2019.03.004 [29] SUI Sai, MA Hua, WANG Jiafu, et al. Absorptive coding metasurface for further radar cross section reduction[J]. Journal of Physics D: Applied Physics, 2018, 51(6): 065603. doi: 10.1088/1361-6463/aaa3be [30] SUI Sai, MA Hua, WANG Jiafu, et al. Synthetic design for a microwave absorber and antireflection to achieve wideband scattering reduction[J]. Journal of Physics D: Applied Physics, 2019, 52(3): 035103. doi: 10.1088/1361-6463/aaeb12 [31] CUI Yanxia, FUNG K H, XU Jun, et al. Ultrabroadband light absorption by a sawtooth anisotropic metamaterial slab[J]. Nano Letters, 2012, 12(3): 1443–1447. doi: 10.1021/nl204118h [32] FU Jiahui, WU Qun, ZHANG Shaoqing, et al. Design of multi-layers absorbers for low frequency applications[C]. 2010 Asia-Pacific International Symposium on Electromagnetic Compatibility, Beijing, China, 2010. doi: 10.1109/APEMC.2010.5475478. [33] SHEN Yang, ZHANG Jieqiu, WANG Jiafu, et al. Multistage dispersion engineering in a three-dimensional plasmonic structure for outstanding broadband absorption[J]. Optical Materials Express, 2019, 9(3): 1539–1550. doi: 10.1364/OME.9.001539 [34] WU C, NEUNER III B, SHVETS G, et al. Large-area wide-angle spectrally selective plasmonic absorber[J]. Physical Review B, 2011, 84(7): 075102. doi: 10.1103/PhysRevB.84.075102 [35] CHENG Yongzhi, GONG Rongzhou, NIE Yan, et al. A wideband metamaterial absorber based on a magnetic resonator loaded with lumped resistors[J]. Chinese Physics B, 2012, 21(12): 127801. doi: 10.1088/1674-1056/21/12/127801 [36] SHEN Yang, ZHANG Jieqiu, WANG Wenjie, et al. Overcoming the pixel-density limit in plasmonic absorbing structure for broadband absorption enhancement[J]. IEEE Antennas and Wireless Propagation Letters, 2019, 18(4): 674–678. doi: 10.1109/LAWP.2019.2900846 [37] ZHU Ruichao, WANG Jiafu, SUI Sai, et al. Wideband absorbing plasmonic structures via profile optimization based on genetic algorithm[J]. Frontiers in Physics, 2020, 8: 231. doi: 10.3389/fphy.2020.00231 [38] NONG Jifu. Global exponential stability of delayed Hopfield neural networks[C]. 2012 International Conference on Computer Science and Information Processing, Xi’an, China, 2012. doi: 10.1109/CSIP.2012.6308827. [39] AN Jinliang, GAO Jia, LEI Jinhui, et al. An improved algorithm for TSP problem solving with Hopfield neural networks[J]. Advanced Materials Research, 2010, 143/144: 538–542. doi: 10.4028/www.scientific.net/AMR.143-144.538 [40] AIYER S V B, NIRANJAN M, and FALLSIDE F. A theoretical investigation into the performance of the Hopfield model[J]. IEEE Transactions on Neural Networks, 1990, 1(2): 204–215. doi: 10.1109/72.80232 [41] ZHU Ruichao, QIU Tianshuo, WANG Jiafu, et al. Metasurface design by a hopfield network: Finding a customized phase response in a broadband[J]. Journal of Physics D: Applied Physics, 2020, 53(41): 415001. doi: 10.1088/1361-6463/ab9785 [42] VAN DEN DRIESSCHE P and ZOU Xingfu. Global attractivity in delayed hopfield neural network models[J]. SIAM Journal on Applied Mathematics, 1998, 58(6): 1878–1890. doi: 10.1137/S0036139997321219 [43] RECH P C. Chaos and hyperchaos in a Hopfield neural network[J]. Neurocomputing, 2011, 74(17): 3361–3364. doi: 10.1016/j.neucom.2011.05.016 [44] SELL D, YANG Jianji, DOSHAY S, et al. Large-angle, multifunctional metagratings based on freeform multimode geometries[J]. Nano Letters, 2017, 17(6): 3752–3757. doi: 10.1021/acs.nanolett.7b01082 [45] FOROUZMAND A and MOSALLAEI H. Composite multilayer shared-aperture nanostructures: A functional multispectral control[J]. ACS Photonics, 2018, 5(4): 1427–1439. doi: 10.1021/acsphotonics.7b01441 [46] TIPRAQSA P, CRASWELL E T, NOBLE A D, et al. Resource integration for multiple benefits: Multifunctionality of integrated farming systems in Northeast Thailand[J]. Agricultural Systems, 2007, 94(3): 694–703. doi: 10.1016/j.agsy.2007.02.009 [47] LING Yonghong, HUANG Lirong, HONG Wei, et al. Polarization-switchable and wavelength-controllable multi-functional metasurface for focusing and surface-plasmon-polariton wave excitation[J]. Optics Express, 2017, 25(24): 29812–29821. doi: 10.1364/OE.25.029812 [48] MAGUID E, YULEVICH I, YANNAI M, et al. Multifunctional interleaved geometric-phase dielectric metasurfaces[J]. Light: Science & Applications, 2017, 6(8): e17027. [49] HUANG Cheng, ZHANG Changlei, YANG Jianing, et al. Reconfigurable metasurface for multifunctional control of electromagnetic waves[J]. Advanced Optical Materials, 2017, 5(22): 1700485. doi: 10.1002/adom.201700485 [50] ZHU Ruichao, QIU Tianshuo, WANG Jiafu, et al. Multiplexing the aperture of a metasurface: Inverse design via deep-learning-forward genetic algorithm[J]. Journal of Physics D: Applied Physics, 2020, 53(45): 455002. doi: 10.1088/1361-6463/aba64f 期刊类型引用(2)
1. 吴强,邓佩佩,陈仁爱,张强辉,安健飞,黄昆,周人,成彬彬. 一种基于太赫兹成像的复杂地形自适应定高方法. 太赫兹科学与电子信息学报. 2024(06): 617-626 . 百度学术
2. 张宇,金潇. 一种前视SAR二维空变误差校正方法. 现代雷达. 2024(10): 1-7 . 百度学术
其他类型引用(1)
-