• Acta Photonica Sinica
  • Vol. 48, Issue 12, 1206002 (2019)
Chun-fu HUANG, An LI, Fang-jun QIN*, and Zhi WANG
Author Affiliations
  • School of Electrical Engineering, Naval University of Engineering, Wuhan 430033, China
  • show less
    DOI: 10.3788/gzxb20194812.1206002 Cite this Article
    Chun-fu HUANG, An LI, Fang-jun QIN, Zhi WANG. Temperature Error Modeling and Real-time Compensation of Fiber Optic Gyroscope Based on PSO-SVR[J]. Acta Photonica Sinica, 2019, 48(12): 1206002 Copy Citation Text show less
    PSO-SVR算法流程图PSO-SVR algorithm flowchart
    Fig. 1. PSO-SVR算法流程图PSO-SVR algorithm flowchart
    温度变化率实时获取Real-time acquisition of temperature change rate
    Fig. 2. 温度变化率实时获取Real-time acquisition of temperature change rate
    FOG温度误差在线补偿FOG temperature error online compensation
    Fig. 3. FOG温度误差在线补偿FOG temperature error online compensation
    最小二乘补偿效果Compensation effect of least squares
    Fig. 4. 最小二乘补偿效果Compensation effect of least squares
    RBF神经网络补偿效果Compensation effect of RBF neural networks
    Fig. 5. RBF神经网络补偿效果Compensation effect of RBF neural networks
    PSO-SVR补偿效果Compensation effect of PSO-SVR
    Fig. 6. PSO-SVR补偿效果Compensation effect of PSO-SVR
    最小二乘法实时补偿效果Real-time compensation effect of least squares
    Fig. 7. 最小二乘法实时补偿效果Real-time compensation effect of least squares
    RBF神经网络实时补偿效果Real-time compensation effect of RBF neural networks
    Fig. 8. RBF神经网络实时补偿效果Real-time compensation effect of RBF neural networks
    PSO-SVR实时补偿效果Real-time compensation effect of PSO-SVR
    Fig. 9. PSO-SVR实时补偿效果Real-time compensation effect of PSO-SVR
    ParametersValue
    Population size n60
    Number of iterations m100
    Learning factor c12
    Learning factor c22
    Table 1. PSO parameters setting
    ParametersεCσ
    Value3.59×10-551.8688.22
    Table 2. Best parameters for SVR obtained by PSO
    SchemesRMSEMaximum error/(°·h-1)
    Before compensation8.09×10-21.28×10-1
    Least squares9.71×10-33.34×10-2
    RBF neural networks5.50×10-32.69×10-2
    PSO-SVR4.28×10-47.40×10-4
    Table 3. Comparison of results of threemodeling methods
    SchemesRMSEMaximumerror/(°·h-1)
    Before compensation6.99×10-21.24×10-1
    Least squares1.55×10-21.00×10-1
    RBF neural networks2.02×10-21.52×10-1
    PSO-SVR1.32×10-26.54×10-2
    Table 4. Comparison of real-time compensation results
    SchemesTime of 10 000 compensation points/sAverage time of per compensation point/s
    Least squares0.151.50×10-5
    RBF neural networks88.608.86×10-3
    PSO-SVR1.111.11×10-4
    Table 5. Operation time
    Chun-fu HUANG, An LI, Fang-jun QIN, Zhi WANG. Temperature Error Modeling and Real-time Compensation of Fiber Optic Gyroscope Based on PSO-SVR[J]. Acta Photonica Sinica, 2019, 48(12): 1206002
    Download Citation