Yuichiro Nishikawa
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Yuichiro Nishikawa has a total of 5 co-authors. Here are the top co-authors by number of shared publications:
- Akira Hirano
Graduate School of Information and Communication Engineering, Tokyo Denki University, Adachi-ku, Tokyo 120-8551, Japan
2 shared publications Akira Hirano
Tokyo Denki University,Information and Communication Engineering,Tokyo,Japan,120-8551
1 shared publicationShota Nishijima
Tokyo Denki University
1 shared publication- Shota Nishijima
Graduate School of Information and Communication Engineering, Tokyo Denki University, Adachi-ku, Tokyo 120-8551, Japan
1 shared publication - Takeshi Naito
Graduate School of Information and Communication Engineering, Tokyo Denki University, Adachi-ku, Tokyo 120-8551, Japan
1 shared publication
Latest Publications
- Open Access
- 17 February 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (4) , 2107
Real-Time Failure Prediction of ROADMs by GAN-Enhanced Machine Learning
We proposed a novel technique for detecting optical filter shift in ROADMs in optical transmission lines by applying machine learning on DP-16QAM constellation data captured just after Analogue-to-Digital Converters (ADCs) in a digital coherent receiver. For this purpose, we implemented Docker container applications in WhiteBox Cassini to acquire the real-time raw digital data. By using the captured data, we generated CNN models for the detections in off-line processing and used them for real-time detections. As preliminary results, we confirmed the successful detection of optical filter shift in real-time with an accuracy of 51 GHz. To enhance the detection accuracy, we challenged ourselves to reproduce digital coherent constellation data by using a Generative Adversarial Network (GAN) for real-time optical filter shift detection for the first time. By utilizing a GAN, we successfully generated clone data based on actual constellation data. By adding the cloned data onto the actually captured data, we successfully enhanced detection sensitivity to as high as 26 GHz. As a result, we reduced the amount of required data for the high detection accuracy by 68% with the help of GAN-supported data augmentation. Furthermore, we confirmed our augmentation method enables the prediction of faults before they occur by enabling high enough detection sensitivity to detect an optical filter shift before degradation of Bit Error Rates (BERs) appears. This demonstrates the potential of GAN-based data augmentation in optimizing the efficiency and precision of optical network impairment sensing by using captured digital coherent optical signal.
- 17 February 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (4) , 2107
Open AccessReal-Time Failure Prediction of ROADMs by GAN-Enhanced Machine Learning
We proposed a novel technique for detecting optical filter shift in ROADMs in optical transmission lines by applying machine learning on DP-16QAM constellation data captured just after Analogue-to-Digital Converters (ADCs) in a digital coherent receiver. For this purpose, we implemented Docker container applications in WhiteBox Cassini to acquire the real-time raw digital data. By using the captured data, we generated CNN models for the detections in off-line processing and used them for real-time detections. As preliminary results, we confirmed the successful detection of optical filter shift in real-time with an accuracy of 51 GHz. To enhance the detection accuracy, we challenged ourselves to reproduce digital coherent constellation data by using a Generative Adversarial Network (GAN) for real-time optical filter shift detection for the first time. By utilizing a GAN, we successfully generated clone data based on actual constellation data. By adding the cloned data onto the actually captured data, we successfully enhanced detection sensitivity to as high as 26 GHz. As a result, we reduced the amount of required data for the high detection accuracy by 68% with the help of GAN-supported data augmentation. Furthermore, we confirmed our augmentation method enables the prediction of faults before they occur by enabling high enough detection sensitivity to detect an optical filter shift before degradation of Bit Error Rates (BERs) appears. This demonstrates the potential of GAN-based data augmentation in optimizing the efficiency and precision of optical network impairment sensing by using captured digital coherent optical signal.
- 1 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEICE Transactions on Communications
- Vol. E106.B (11) , 1065-1073
Real-Time Detection of Fiber Bending and/or Optical Filter Shift by Machine-Learning of Tapped Raw Digital Coherent Optical Signals
We have proposed autonomous network diagnosis platform for operation of future large capacity and virtualized network, including 5G and beyond 5G services. As for the one candidate of information collection and analyzing function blocks in the platform, we proposed novel optical sensing techniques that utilized tapped raw signal data acquired from digital coherent optical receivers. The raw signal data is captured before various digital signal processing for demodulation. Therefore, it contains various waveform deformation and/or noise as it experiences through transmission fibers. In this paper, we examined to detect two possible failures in transmission lines including fiber bending and optical filter shift by analyzing the above-mentioned raw signal data with the help of machine learning. For the purpose, we have implemented Docker container applications in White Box Cassini to acquire real-time raw signal data. We generated CNN model for the detections in off-line processing and used them for real-time detections. We have confirmed successful detection of optical fiber bend and/or optical filter shift in real-time with high accuracy. Also, we evaluated their tolerance against ASE noise and invented novel approach to improve detection accuracy. In addition to that, we succeeded to detect them even in the situation of simultaneous occurrence of those failures.
- 1 November 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEICE Transactions on Communications
- Vol. E106.B (11) , 1065-1073
Real-Time Detection of Fiber Bending and/or Optical Filter Shift by Machine-Learning of Tapped Raw Digital Coherent Optical Signals
We have proposed autonomous network diagnosis platform for operation of future large capacity and virtualized network, including 5G and beyond 5G services. As for the one candidate of information collection and analyzing function blocks in the platform, we proposed novel optical sensing techniques that utilized tapped raw signal data acquired from digital coherent optical receivers. The raw signal data is captured before various digital signal processing for demodulation. Therefore, it contains various waveform deformation and/or noise as it experiences through transmission fibers. In this paper, we examined to detect two possible failures in transmission lines including fiber bending and optical filter shift by analyzing the above-mentioned raw signal data with the help of machine learning. For the purpose, we have implemented Docker container applications in White Box Cassini to acquire real-time raw signal data. We generated CNN model for the detections in off-line processing and used them for real-time detections. We have confirmed successful detection of optical fiber bend and/or optical filter shift in real-time with high accuracy. Also, we evaluated their tolerance against ASE noise and invented novel approach to improve detection accuracy. In addition to that, we succeeded to detect them even in the situation of simultaneous occurrence of those failures.
- 3 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-3
Enhanced Accuracy in Optical Fiber Bending Detection with Machine Learning against Accumulated ASE Noise
We have enhanced accuracy of optical fiber bending detection with machine-learning against accumulated ASE noise. Up to 99 of accuracy is obtained by incorporating post-demodulation data with pre-demodulation data in digital coherent optical receiver.
- 3 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-3
Enhanced Accuracy in Optical Fiber Bending Detection with Machine Learning against Accumulated ASE Noise
We have enhanced accuracy of optical fiber bending detection with machine-learning against accumulated ASE noise. Up to 99 of accuracy is obtained by incorporating post-demodulation data with pre-demodulation data in digital coherent optical receiver.