No abstract

This paper presents a novel and general distributed acoustic sensing (DAS) signal recognition framework aimed at real-time detection and classification of intrusion in the space-time domain. The framework is based on the combination of a convolution neural network (CNN) and a long short-term memory network (LSTM). The convolutional structure extracts the spatial features from multi-channel signals of the DAS system, while the LSTM network analyzes the temporal relationships over time. The framework can be deployed on high-speed railways for real-time intrusion threat detection, which is one of the most urgent and challenging problems that needs to be resolved as there is an increasing demand for high detection and low false alarm rates, and short response time. The alarm sensitivity and specificity of the framework are controlled by user-set parameters. A real field experiment is conducted in a strong background noise scenario and an intrusion threat detection rate of 85.6%, with only 8.0% false alarm rate is achieved. For threat classification, the average threat detection rate is 69.3%, and the average false alarm rate is 13.2%. Owing to the high detection accuracy of the framework, the average detection response time is shortened to 8.25 s.

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.

No abstract

No abstract