• Chinese Optics Letters
  • Vol. 18, Issue 11, 111404 (2020)
Shanshan Li1, Qi Zhang1、2、3、*, Xiangjun Xin1、2、3, Ran Gao4, Sitong Zhou1, Ying Tao5, Yufei Shen6, Huan Chang7, Qinghua Tian1、2、3, Feng Tian1、2、3, Yongjun Wang1、2、3, and Leijing Yang1、2、3
Author Affiliations
  • 1School of Electronic Engineering, Beijing University of Posts and Telecommunications (BUPT), Beijing 100876, China
  • 2Beijing Key Laboratory of Space-round Interconnection and Convergence, BUPT, Beijing 100876, China
  • 3State Key Laboratory of Information Photonics and Optical Communications, BUPT, Beijing 100876, China
  • 4The Advanced Research Institute of Multidisciplinary Science, Beijing Institute of Technology, Beijing 100081, China
  • 5China Academy of Space Technology, Beijing 100094, China
  • 6China Satellite Communication Co., Ltd., Beijing 100048, China
  • 7School of Information and Electronics, Beijing Institute of Technology, Beijing 100081, China
  • show less
    DOI: 10.3788/COL202018.111404 Cite this Article Set citation alerts
    Shanshan Li, Qi Zhang, Xiangjun Xin, Ran Gao, Sitong Zhou, Ying Tao, Yufei Shen, Huan Chang, Qinghua Tian, Feng Tian, Yongjun Wang, Leijing Yang. No prior recognition method of modulation mode by partition-fractal and SVM learning method[J]. Chinese Optics Letters, 2020, 18(11): 111404 Copy Citation Text show less
    System scheme of AMC implementation in satellite communication system.
    Fig. 1. System scheme of AMC implementation in satellite communication system.
    Description of the partition-fractal method.
    Fig. 2. Description of the partition-fractal method.
    Calculation process of constellation diagram feature matrix.
    Fig. 3. Calculation process of constellation diagram feature matrix.
    Comparison diagram of classifier performance under different learning algorithms.
    Fig. 4. Comparison diagram of classifier performance under different learning algorithms.
    Classification accuracy in different modulation modes (SNR, in dB). (a) SVM; (b) bagging; (c) KNN; (d) classification tree; (e) AdaBoost.
    Fig. 5. Classification accuracy in different modulation modes (SNR, in dB). (a) SVM; (b) bagging; (c) KNN; (d) classification tree; (e) AdaBoost.
    Symbol rate1200 bit/s
    Sampling frequency4800 Hz
    Frequencies separation5 Hz
    Signal duration1 s
    Signal selectable signal-to-noise ratio range020dB
    Modulation modesBASK, BPSK, BFSK, QASK, QPSK, QFSK, 8ASK, 8PSK, 8FSK
    Number of constellations18,900
    Constellations size512×512pixels
    Training and test ratio7:3
    Table 1. Simulation Parameters
    SNR/dB146121620
    Accuracy0.95190.96670.97410.97780.99261
    Table 2. Accuracy of Classifier under Different SNR
    Shanshan Li, Qi Zhang, Xiangjun Xin, Ran Gao, Sitong Zhou, Ying Tao, Yufei Shen, Huan Chang, Qinghua Tian, Feng Tian, Yongjun Wang, Leijing Yang. No prior recognition method of modulation mode by partition-fractal and SVM learning method[J]. Chinese Optics Letters, 2020, 18(11): 111404
    Download Citation