Shape Coordinate Prediction of Multi-Core Optical Fiber Based on CNN-LSTM Network and Attention Mechanism

Jianjian Wang; Rui Tang

doi:10.3788/AOS241474

Acta Optica Sinica, Volume. 45, Issue 1, 0106003(2025)

Shape Coordinate Prediction of Multi-Core Optical Fiber Based on CNN-LSTM Network and Attention Mechanism

Jianjian Wang^1,2,3 and Rui Tang^1、*

¹Department of Electronic & Communication Engineering, North China Electric Power University, Baoding 071003, Hebei , China

²Hebei Key Laboratory of Power Internet of Things Technology, North China Electric Power University, Baoding 071003, Hebei , China

³Baoding Key Laboratory of Optical Fiber Sensing and Optical Communication Technology, North China Electric Power University, Baoding 071003, Hebei , China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(11)

Fig. 1. Geometric diagrams of bent optical fiber. (a) Bent deformation; (b) sensing cross section

Download full size

Fig. 2. Multi-core optical fiber shape sensing process

Download full size

Fig. 3. Schematic diagram of coordinate prediction based on neural network

Download full size

Fig. 4. Finite element model of multi-core optical fiber

Download full size

Fig. 5. Flowchart of shape coordinate prediction network training

Download full size

Fig. 6. Optical fiber shape sensing experimental system

Download full size

Fig. 7. Comparisons of curve coordinates with a curvature radius of 400 mm. (a) Coordinates of horizontal axis; (b) coordinates of vertical axis

Download full size

Fig. 8. Comparisons of curve coordinates with a curvature radius of 700 mm. (a) Coordinates of horizontal axis; (b) coordinates of vertical axis

Download full size

Fig. 9. Shape sensing results of multi-core optical fiber. (a) Curve with a curvature radius of 400 mm; (b) curve with a curvature radius of 700 mm

Download full size

Table 1. Parameters setting in LSTM
View table
Table 1. Parameters setting in LSTM
Parameter Setting
Input_size 5
Hidden_size 128
Num_layer 2
Bias True
Batch_first False
Dropout 0.4
Bidirectional False

Table 2. Errors of shape sensing results
View table
Table 2. Errors of shape sensing results
Method 400 mm 700 mm
RMSE /mm MAE /mm RMSE /mm MAE /mm
computation 3.6451 2.4287 2.5019 0.9143
LSTM 3.1462 1.9334 2.0742 0.8849
CNN+LSTM 2.9249 1.8558 1.7571 0.7316
MODEL 2.1244 1.6111 1.5739 0.6919

Tools

Get Citation

Copy Citation Text

Jianjian Wang, Rui Tang. Shape Coordinate Prediction of Multi-Core Optical Fiber Based on CNN-LSTM Network and Attention Mechanism[J]. Acta Optica Sinica, 2025, 45(1): 0106003

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites