Acta Optica Sinica, Volume. 42, Issue 4, 0406004(2022)
Deep Learning-Based Recognition of Modes and Mode Groups in Multimode Optical Fiber Communication System
Fig. 3. Simulation results of recognition of different linear polarization modes under different noise intensities. (a) Optical field of LP11a mode; (b) recognition rate of LP11a mode; (c) variation of multiclass logarithmic loss function; (d) mode prediction result of each mode of recognition model
Fig. 4. Recognition results of linear polarization mode groups obtained by recognition model with noise intensity of 20%. (a) Recognition rate; (b) multiclass logarithmic loss function
Fig. 5. Intensity profiles of multimode superimposed light field in mode group. (a) Simulated light field intensity distributions; (b) intensity distributions of output optical field of multimode fiber
Fig. 7. Experimental results of mode group classification and recognition when arrays of light detectors of different sizes are used to receive light field. (a) Variation of recognition rate; (b) variation of multiclass logarithmic loss function
Get Citation
Copy Citation Text
Jinkun Hu, Xiaojie Guo, Jianping Li, Ou Xu, Meng Xiang, Di Peng, Songnian Fu, Yuwen Qin. Deep Learning-Based Recognition of Modes and Mode Groups in Multimode Optical Fiber Communication System[J]. Acta Optica Sinica, 2022, 42(4): 0406004
Category: Fiber Optics and Optical Communications
Received: Jun. 24, 2021
Accepted: Sep. 6, 2021
Published Online: Jan. 29, 2022
The Author Email: Li Jianping (jianping@gdut.edu.cn)