本實驗為馬赫詹德干涉儀(Mach-Zehnder Interferometer，MZI)搭配自製窄線寬雷射為光源之分布式入侵感測系統，透過對於入侵感測系統做出三種入侵行為，分別為環境干擾、輕拍及重拍。從入侵感測系統取得不同事件的一維時域訊號波形，透過設立重拍之閾值將事件做分類，其閾值Signal Amplitude為2，Level Crossing為12，Frequency Ratio為0.15，並透過一維時域訊號波形轉換為二維循環圖(Recurrence Plot)之訊號預處理，讓不同事件的時域訊號波形具有明顯的差異性。
將不同事件遞歸圖形做事件標記，丟入卷積神經網絡(Convolutional Neural Network，CNN)中進行訓練及更新參數並改良測試模型以提升模型的事件判別準確率，經過數個模型去做比較改良，實驗後最終模型的準確率達到86.66 %。一張圖取樣時間為10 ms，每100張為一組的1 s數據組，透過判別數據組中成分比例來判別數據組為何種事件以及檢測模型準確性，在模型最後設立成分比例閾值後來進行數據組的判別，經過實驗後模型在數據組上具有不錯的判別表現，環境干擾、輕拍與重拍的數據組準確率都有達到九成以上。

This experiment involves a Mach-Zehnder Interferometer (MZI) integrated with a self-made narrow linewidth laser as the light source for a distributed intrusion detection system. Three types of signals disturbance arising from different behaviors, namely environment signal, tap signal, and knock signal, are generated for study in the intrusion detection system. One-dimensional time-domain signal waveforms are obtained from the intrusion detection system for different events. A threshold of knocking to classify the event for intrusion is set with Signal Amplitude at 2, Level Crossing at 12, and Frequency Ratio at 0.15. The preprocessing step is performed to transform the time-domain signal waveforms into two-dimensional recurrence plots, enhancing the distinguishability between different event signals.
Different events are labeled based on the recurrence plots and fed into a Convolutional Neural Network (CNN) for training and parameter updating. The CNN model is continuously improved to enhance the accuracy of identifying events. After comparing and refining several models through experiments, the final model achieves an accuracy of 86.66%. Each image has a sampling time of 10 ms, and a data set consists of 100 images taken within a 1 s time frame. The classification of data sets into different types of events and the evaluation of model accuracy are based on the proportions of different event components. After setting criterion for each type of event at the end of the model, the data sets are classified in the experiment. The model shows good performance in data set classification, with accuracy exceeding 90% for identifying environment disturbance, tapping, and knocking data sets.

[1] K. C. Kao and G. A. Hockham, “Dielectric-fiber surface waveguides for optical frequencies,” Proceedings of the IEEE. Institute of Electrical and Electronics Engineers, vol. 113, pp. 1151–1158, 1966.
[2] J. E. Antonio-Lopez, Z. S. Eznaveh, P. LiKamWa, A. Schülzgen, and R. Amezcua-Correa, “Multicore fiber sensor for high-temperature applications up to 1000°C,” Optics Letters, vol. 39, no. 15, pp. 4309-4312, 2014.
[3] Y. Geng, X. Li, X. Tan, Y. Deng, and X. Hong, “Compact and ultrasensitive temperature sensor with a fully liquid-filled photonic crystal fiber Mach–Zehnder interferometer,” IEEE Sensors Journal, vol. 14, no. 1, pp.167-170, 2014.
[4] F. Zhu, Y. Zhang, and L. Xia, “Improved Φ-OTDR sensing system for high-precision dynamic strain measurement based on ultra-weak fiber Bragg grating array,” Journal of Lightwave Technology, vol. 33, no. 23, pp. 4775-4780, 2015.
[5] X. Zhang and W. Peng, “Bent fiber interferometer,” Journal of Lightwave Technology, vol. 33, no. 15, pp. 3351-3356, 2015.
[6] M. Deng, Y. Zhao, F. Yin, and T. Zhu, “Interferometric fiber-optic tilt sensor exploiting taper and lateral-offset fusing splicing,” IEEE Photonics Technology Letters, vol. 28, no. 20, pp. 2225-2228, 2016.
[7] J. Kong, X. Ouyang, A. Zhou, and L. Yuan, “Highly sensitive directional bending sensor based on eccentric core fiber Mach–Zehnder modal interferometer,” IEEE Sensors Journal, vol. 16, no. 18, pp. 6899-6902, 2016.
[8] R. Gao, Y. Jiang, and S. Abdelaziz, “All-fiber magnetic field sensors based on magnetic fluid-filled photonic crystal fibers,” Optics Letters, vol. 38, no. 9, pp. 1539-1541, 2013.

[9] Y. Zheng, X. Dong, C. C. Chan, P. P. Shum, and H. Su, “Optical fiber magnetic field sensor based on magnetic fluid and microfiber mode interferometer,” Optics Communications, vol. 336, pp. 5-8, 2015.
[10] F. Shi, X. Bai, F. Wang, F. Pang, S. Pu, and X. Zeng, “All-fiber magnetic field sensor based on hollow optical fiber and magnetic fluid,” IEEE Sensors Journal, vol. 17, no. 3, pp. 619-622, 2017.
[11] C. Waltermann, J. Koch, and M. Angelmahr, “Femtosecond laser aided processing of optical sensor fibers for 3D medical navigation and tracking (FiberNavi),” 23rd International Conference on Optical Fiber Sensors. International Society for Optics and Photonics, Proceedings of SPIE, vol. 9157, pp. 91577G1-91577G-4, 2014.
[12] M. Hassan, E. Gonzalez, V. Hitchins, and I. Ilev, “Detecting bacteria contamination on medical device surfaces using an integrated fiber-optic mid-infrared spectroscopy sensing method,” Sensors and Actuators B: Chemical, vol. 231, pp. 646-654, 2016.
[13] J. Gao, Z. Jiang, Y. Zhao, L. Zhu, and G. Zhao, "Full distributed fiber optical sensor for intrusion detection in application to buried pipelines," Chinese Optics Letters, vol. 3, Issue 11, pp. 633-635 2005.
[14] A. Catalano, F. A. Bruno, C. Galliano, M. Pisco, G. V. Persiano, A. Cutolo, A. Cusano, "An optical fiber intrusion detection system for railway security," Sensors and Actuators A: Physical, vol. 253, pp. 91–100 2017.
[15] Y. Lu, T. Zhu, L. Chen, and X. Bao, “Distributed vibration sensor based on coherent detection of phase-OTDR,” Journal of Lightwave Technology, vol. 28, no. 22, pp. 3243–3249, 2010.
[16] Paul R. Hoffman and Mark G. Kuzyk, “Position determination of an acoustic burst along a Sagnac interferometer,” Journal of Lightwave Technology, vol. 22, no. 2, pp. 494–498, 2004.
[17] C. Ma, T. Liu, K. Liu, J. Jiang, Z. Ding, L. Pan, and M. Tian, “Long-range distributed fiber vibration sensor using an asymmetric dual Mach–Zehnder interferometers,” Journal of Lightwave Technology, vol. 34, no. 9, pp. 2235-2239, 2016.
[18] K. Liu et al., "A high-efficiency multiple events discrimination method in optical fiber perimeter security system," Journal of Lightwave Technology, vol. 33, no. 23, pp. 4885-4890, 1 Dec.1, 2015
[19] C. Lyu, Z. Huo, X. Cheng, J. Jiang, A. Alimasi and H. Liu, "Distributed optical fiber sensing intrusion pattern recognition based on GAF and CNN," Journal of Lightwave Technology, vol. 38, no. 15, pp. 4174-4182, 1 Aug.1, 2020.
[20] J.P. Eckmann, S. Kamphorst, D. Ruelle, "Recurrence plots of dynamical systems," Europhysics Letters, vol. 4, pp. 973-977, 1987.
[21] [1] C. Lyu, J. Jiang, Z. Huo, "Open-set events identification based on deep metric-learning for DMZI perimeter system," IEEE Sensors Journal, vol. 20, no. 22, pp. 13620-13628, 2020.
[22] N. Marwan, M. Carmenromano, M. Thiel, and J. Kurths, “Recurrence plots for the analysis of complex systems,” Phys. Rep., vol. 438, nos. 5–6, pp. 237–329, Jan. 2007.
[23] D.H. Hubel, T.N. Wiesel, "Receptive fields and functional architecture of monkey striate cortex," Journal of physiology, vol. 195, pp. 215-43, 1968.
[24] K. Fukushima, "Cognitron: A self-organizing multilayered neural network," Biological cybernetics, vol. 20, pp. 121-136, 1975.
[25] Nair, Vinod and Hinton, Geoffrey, "Rectified linear units improve restricted boltzmann machines vinod nair," Proceedings of ICML, vol. 27, pp. 807-814, 2010.
[26] Christopher M. Bishop and Nasser M. Nasrabadi, "Pattern recognition and machine learning," J. Electronic Imaging, vol. 4, p. 0949901, 2006.
[27] H. Yang, "Multiscale recurrence quantification analysis of spatial vectorcardiogram (VCG) signals," IEEE Transactions on Biomedical Engineering, vol. 58, no. 2, pp. 339-347, 2011.
[28] Y. Chen and H. Yang, "Multiscale recurrence analysis of long-term nonlinear and nonstationary time series," Chaos, Solitons & Fractals, vol. 45, no. 7, pp. 978-987, 2012.
[29] 謝承諭，「以光纖雷射用於光纖雙馬赫詹德干涉儀入侵感測定位系統：平均法改善定位解析度」，國立清華大學光電工程研究所碩士班碩士論文，民國一百零七年十一月。
[30] M. Claesen and B. De Moor, "Hyperparameter search in machine learning," 2015.
[31] D. Kingma and J. Ba, "Adam: A method for stochastic optimization," International Conference on Learning Representations, 2014.
[32] E. Hinton, Geoffrey, et al. "System and method for addressing overfitting in a neural network." U.S. Patent, No. 9,406,017, 2016.
[33] https://blog.paperspace.com/busting-the-myths-about-batch-normalization/
[34] S. Ioffe and C. Szegedy, "Batch normalization: Accelerating deep network training by reducing internal covariate shift," International conference on machine learning, pp. 448-456, 2015.
[35] P. Ma, K. Liu, J. Jiang, Z. Li, P. Li, and T. Liu, "Probabilistic event discrimination algorithm for fiber optic perimeter security Systems," Journal of Lightwave Technology, vol. 36, no. 11, pp. 2069-2075, 2018.
[36] V. Sitzmann, J. Martel, A. Bergman, D. Lindell, and G. Wetzstein, "Implicit neural representations with periodic activation functions,", 2020.