Label-Free Anomaly Detection Using Distributed Optical Fiber Acoustic Sensing
Deep learning anomaly detection is important in distributed optical fiber acoustic sensing (DAS). However, anomaly detection is more challenging than traditional learning tasks, due to the scarcity of true-positive data and the vast imbalance and irregularity within datasets. Furthermore, it is impo...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-04-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/23/8/4094 |
| _version_ | 1797603482784497664 |
|---|---|
| author | Yuyuan Xie Maoning Wang Yuzhong Zhong Lin Deng Jianwei Zhang |
| author_facet | Yuyuan Xie Maoning Wang Yuzhong Zhong Lin Deng Jianwei Zhang |
| author_sort | Yuyuan Xie |
| collection | DOAJ |
| description | Deep learning anomaly detection is important in distributed optical fiber acoustic sensing (DAS). However, anomaly detection is more challenging than traditional learning tasks, due to the scarcity of true-positive data and the vast imbalance and irregularity within datasets. Furthermore, it is impossible to catalog all types of anomalies, therefore, the direct application of supervised learning is deficient. To overcome these problems, an unsupervised deep learning method that only learns the normal data features from ordinary events is proposed. First, a convolutional autoencoder is used to extract DAS signal features. A clustering algorithm then locates the feature center of the normal data, and the distance to the new signal is used to determine whether it is an anomaly. The efficacy of the proposed method was evaluated in a real high-speed rail intrusion scenario, and considered all behaviors that may threaten the normal operation of high-speed trains as abnormal. The results show that the threat detection rate of this method reaches 91.5%, which is 5.9% higher than that of the state-of-the-art supervised network and, at 7.2%, the false alarm rate is 0.8% lower than the supervised network. Moreover, using a shallow autoencoder reduces the parameters to 1.34 K, which is significantly lower than the 79.55 K of the state-of-the-art supervised network. |
| first_indexed | 2024-03-11T04:32:46Z |
| format | Article |
| id | doaj.art-257d42b7c36446cfa3f17bd9b8aae4fd |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-11T04:32:46Z |
| publishDate | 2023-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-257d42b7c36446cfa3f17bd9b8aae4fd2023-11-17T21:18:59ZengMDPI AGSensors1424-82202023-04-01238409410.3390/s23084094Label-Free Anomaly Detection Using Distributed Optical Fiber Acoustic SensingYuyuan Xie0Maoning Wang1Yuzhong Zhong2Lin Deng3Jianwei Zhang4Sichuan University National Key Laboratory of Fundamental Science on Synthetic Vision, Sichuan University, Chengdu 610064, ChinaCollege of Electrical Engineering, Sichuan University, Chengdu 610065, ChinaCollege of Electrical Engineering, Sichuan University, Chengdu 610065, ChinaSichuan University National Key Laboratory of Fundamental Science on Synthetic Vision, Sichuan University, Chengdu 610064, ChinaSichuan University National Key Laboratory of Fundamental Science on Synthetic Vision, Sichuan University, Chengdu 610064, ChinaDeep learning anomaly detection is important in distributed optical fiber acoustic sensing (DAS). However, anomaly detection is more challenging than traditional learning tasks, due to the scarcity of true-positive data and the vast imbalance and irregularity within datasets. Furthermore, it is impossible to catalog all types of anomalies, therefore, the direct application of supervised learning is deficient. To overcome these problems, an unsupervised deep learning method that only learns the normal data features from ordinary events is proposed. First, a convolutional autoencoder is used to extract DAS signal features. A clustering algorithm then locates the feature center of the normal data, and the distance to the new signal is used to determine whether it is an anomaly. The efficacy of the proposed method was evaluated in a real high-speed rail intrusion scenario, and considered all behaviors that may threaten the normal operation of high-speed trains as abnormal. The results show that the threat detection rate of this method reaches 91.5%, which is 5.9% higher than that of the state-of-the-art supervised network and, at 7.2%, the false alarm rate is 0.8% lower than the supervised network. Moreover, using a shallow autoencoder reduces the parameters to 1.34 K, which is significantly lower than the 79.55 K of the state-of-the-art supervised network.https://www.mdpi.com/1424-8220/23/8/4094deep learningdistributed optical fiber acoustic sensingunsupervised learning |
| spellingShingle | Yuyuan Xie Maoning Wang Yuzhong Zhong Lin Deng Jianwei Zhang Label-Free Anomaly Detection Using Distributed Optical Fiber Acoustic Sensing Sensors deep learning distributed optical fiber acoustic sensing unsupervised learning |
| title | Label-Free Anomaly Detection Using Distributed Optical Fiber Acoustic Sensing |
| title_full | Label-Free Anomaly Detection Using Distributed Optical Fiber Acoustic Sensing |
| title_fullStr | Label-Free Anomaly Detection Using Distributed Optical Fiber Acoustic Sensing |
| title_full_unstemmed | Label-Free Anomaly Detection Using Distributed Optical Fiber Acoustic Sensing |
| title_short | Label-Free Anomaly Detection Using Distributed Optical Fiber Acoustic Sensing |
| title_sort | label free anomaly detection using distributed optical fiber acoustic sensing |
| topic | deep learning distributed optical fiber acoustic sensing unsupervised learning |
| url | https://www.mdpi.com/1424-8220/23/8/4094 |
| work_keys_str_mv | AT yuyuanxie labelfreeanomalydetectionusingdistributedopticalfiberacousticsensing AT maoningwang labelfreeanomalydetectionusingdistributedopticalfiberacousticsensing AT yuzhongzhong labelfreeanomalydetectionusingdistributedopticalfiberacousticsensing AT lindeng labelfreeanomalydetectionusingdistributedopticalfiberacousticsensing AT jianweizhang labelfreeanomalydetectionusingdistributedopticalfiberacousticsensing |