• Laser & Optoelectronics Progress
  • Vol. 59, Issue 13, 1323001 (2022)
Jingchang Nan, Youyi Du*, Minghuan Wang, and Mingming Gao
Author Affiliations
  • School of Electronics and Information Engineering, Liaoning Technical University, Huludao 125105, Liaoning , China
  • show less
    DOI: 10.3788/LOP202259.1323001 Cite this Article Set citation alerts
    Jingchang Nan, Youyi Du, Minghuan Wang, Mingming Gao. Deep Learning Architecture and Neural Network Optimization of Ultra-Wideband Antenna Modeling[J]. Laser & Optoelectronics Progress, 2022, 59(13): 1323001 Copy Citation Text show less
    Structure block diagram of multi-layer perceptron and hidden neurons
    Fig. 1. Structure block diagram of multi-layer perceptron and hidden neurons
    Adam optimizer algorithm framework flow
    Fig. 2. Adam optimizer algorithm framework flow
    DMLP neural network architecture
    Fig. 3. DMLP neural network architecture
    Structure diagram of ultra-wideband stepped microstrip monopole antenna. (a) Front antenna; (b) back antenna
    Fig. 4. Structure diagram of ultra-wideband stepped microstrip monopole antenna. (a) Front antenna; (b) back antenna
    HFSS simulation result of S11 characteristic curve of ultra-wideband stepped microstrip monopole antenna
    Fig. 5. HFSS simulation result of S11 characteristic curve of ultra-wideband stepped microstrip monopole antenna
    Flow chart of DMLP neural network
    Fig. 6. Flow chart of DMLP neural network
    Optimization process of DMLP network, RBF network, and MLP network
    Fig. 7. Optimization process of DMLP network, RBF network, and MLP network
    Prediction curves of three kinds of networks and their HFSS simulation curves fitting in original data
    Fig. 8. Prediction curves of three kinds of networks and their HFSS simulation curves fitting in original data
    Prediction curves of three kinds of networks and their HFSS simulation curves fitting in interpolation data
    Fig. 9. Prediction curves of three kinds of networks and their HFSS simulation curves fitting in interpolation data
    ParameterValueParameterValueParameterValue

    L

    W1

    L1

    LP

    LG

    22

    1.5

    3

    12

    8.7

    W

    W2

    L2

    LS

    LD

    13

    1.3

    5

    4.216

    3.6

    H

    W3

    L3

    WS

    WD

    0.8

    1

    2

    3.5

    1.295

    Table 1. Specific geometric parameters of antenna
    Geometric parameterMinimum valueMaximum valueSampling step sizeNumber of samples

    W1 /mm

    W2 /mm

    W3 /mm

    LD /mm

    WD /mm

    LG /mm

    WS /mm

    LS /mm

    Frequency /GHz

    1.4

    1.20

    0.9

    3.6

    1.270

    8.6

    3.4

    4.108

    2

    1.6

    1.30

    1.0

    3.7

    1.295

    8.7

    3.5

    4.216

    24

    0.1

    0.05

    0.1

    0.1

    0.025

    0.1

    0.1

    0.108

    0.1

    3

    3

    2

    2

    2

    2

    2

    2

    221

    Table 2. Sampling point variables
    Neural network modelMAERMSEMREMaximum relative error

    DMLP

    MLP

    RBF

    0.2860

    1.4692

    1.5236

    0.3948

    2.8208

    2.9171

    1.6701

    7.0359

    7.3067

    1.877

    13.300

    13.780

    Table 3. Performance comparison of three networks
    Jingchang Nan, Youyi Du, Minghuan Wang, Mingming Gao. Deep Learning Architecture and Neural Network Optimization of Ultra-Wideband Antenna Modeling[J]. Laser & Optoelectronics Progress, 2022, 59(13): 1323001
    Download Citation