• Laser & Optoelectronics Progress
  • Vol. 59, Issue 17, 1706001 (2022)
Tonghao Zhang, Xudong Wang*, and Nan Wu
Author Affiliations
  • Information Science and Technology College, Dalian Maritime University, Dalian 116026, Liaoning , China
  • show less
    DOI: 10.3788/LOP202259.1706001 Cite this Article Set citation alerts
    Tonghao Zhang, Xudong Wang, Nan Wu. CNN-Based Hybrid QAM-PPM Modulation End-to-End Communication System[J]. Laser & Optoelectronics Progress, 2022, 59(17): 1706001 Copy Citation Text show less
    Block diagram of QAM-PPM hybrid modulation system under CNN-AE architecture
    Fig. 1. Block diagram of QAM-PPM hybrid modulation system under CNN-AE architecture
    Specific structure of dot_sigma layer
    Fig. 2. Specific structure of dot_sigma layer
    Characteristics of sigmoid function with different δ values
    Fig. 3. Characteristics of sigmoid function with different δ values
    Comparison of SER performance at different orders. (a) M=4 and 8; (b) M=16 and 32; (c) M=64
    Fig. 4. Comparison of SER performance at different orders. (a) M=4 and 8; (b) M=16 and 32; (c) M=64
    Comparison of SER performance when N=8 and 10
    Fig. 5. Comparison of SER performance when N=8 and 10
    Comparison of SER performance when M=4, 8, 16, 32, and 64
    Fig. 6. Comparison of SER performance when M=4, 8, 16, 32, and 64
    Comparison of SER performance when N=8 and 10
    Fig. 7. Comparison of SER performance when N=8 and 10
    Loss functions of training set and validation set for two modulation schemes under AWGN channel. (a) M=4 and N=8; (b) M=64 and N=4
    Fig. 8. Loss functions of training set and validation set for two modulation schemes under AWGN channel. (a) M=4 and N=8; (b) M=64 and N=4
    Loss functions of training set and validation set for two modulation schemes under Rayleigh channel. (a) M=4 and N=8; (b) M=64 and N=4
    Fig. 9. Loss functions of training set and validation set for two modulation schemes under Rayleigh channel. (a) M=4 and N=8; (b) M=64 and N=4
    Comparison of SER performance of AWGN channel CNN and DNN
    Fig. 10. Comparison of SER performance of AWGN channel CNN and DNN
    Comparison of SER performance of Rayleigh channel CNN and DNN
    Fig. 11. Comparison of SER performance of Rayleigh channel CNN and DNN
    Layer typeActivation functionOutput dimension
    InputNoneBL2k1+2k2
    s1NoneBL2k1
    s2NoneBL2k2
    E1_Conv1DeluBL,256)

    E1_Conv1D

    E1_Conv1D

    elu

    elu

    BL,256)

    BL,256)

    E1_Conv1DlinearBL,2)
    E2_Conv1DreluBL,64)
    E2_Conv1DreluBL,64)
    E2_Conv1DlinearBLN
    dot_sigmasigmoidBLN
    ConcatenateNoneBL,2+N
    QAM_PPMNoneBLN
    NormalizationNoneBLN
    ChannelNoneBN×L,2)
    Conv1DeluBL,256)
    Conv1DeluBL,256)
    Conv1DeluBL,256)
    Conv1DeluBL,256)
    Conv1DsoftmaxBL,2k
    Table 1. Parameters for autoencoder network structure
    SystemLearned modulationTraining parameterTraining time /s
    DNN-AE-QAM-PPM4QAM-4PPM34576214067
    16QAM-4PPM36117014131
    64QAM-4PPM42280214580
    CNN-AE-QAM-PPM4QAM-4PPM3457627201
    16QAM-4PPM3611707333
    64QAM-4PPM4228027556
    Table 2. System complexity comparison between DNN and CNN
    Tonghao Zhang, Xudong Wang, Nan Wu. CNN-Based Hybrid QAM-PPM Modulation End-to-End Communication System[J]. Laser & Optoelectronics Progress, 2022, 59(17): 1706001
    Download Citation