Intrusion detection for high-speed railways based on unsupervised anomaly detection models

Wang, Yao; Yu, Zujun; Zhu, Liqiang

doi:10.1007/s10489-022-03911-8

Cited by 11 publications

(2 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…these AE-based methods fail to detect small obstacles or those with colors common in the railway environment. Wang et al [43] similarly train an AE, but detect anomalies by directly analyzing the distribution in the latent space and solely utilize the reconstruction for localizing detected anomalies.…”

Section: B Visual Obstacle Detection On Railwaysmentioning

confidence: 99%

Local and Global Information in Obstacle Detection on Railway Tracks

Brucker,

Cramariuc,

Von Einem

et al. 2023

2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

View full text Add to dashboard Cite

Reliable obstacle detection on railways could help prevent collisions that result in injuries and potentially damage or derail the train. Unfortunately, generic object detectors do not have enough classes to account for all possible scenarios, and datasets featuring objects on railways are challenging to obtain. We propose utilizing a shallow network to learn railway segmentation from normal railway images. The limited receptive field of the network prevents overconfident predictions and allows the network to focus on the locally very distinct and repetitive patterns of the railway environment. Additionally, we explore the controlled inclusion of global information by learning to hallucinate obstacle-free images. We evaluate our method on a custom dataset featuring railway images with artificially augmented obstacles. Our proposed method outperforms other learning-based baseline methods.

show abstract

Section: B Visual Obstacle Detection On Railwaysmentioning

confidence: 99%

Local and Global Information in Obstacle Detection on Railway Tracks

Brucker,

Cramariuc,

Von Einem

et al. 2023

2023 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

View full text Add to dashboard Cite

show abstract

“…The generational transition has resulted in the creation of increasingly complex security function methods. One of the most effective safety measures is network anomaly detection and prediction, which has gained popularity [4]. The linear models known as vector autoregressive (VAR) designs show the relationship between the lags of a particular variable, allowing for the forecasting or prediction of future values.…”

Section: Introductionmentioning

confidence: 99%

A Comparative Analysis of IoT based Network Anomaly Detection and Prediction Using Vector Autoregressive Models

Cho,

Cho

2024

JMC

View full text Add to dashboard Cite

This research provides a comparative analysis of the use of Vector Autoregressive models for network anomaly detection and prediction. It starts by giving a brief overview of the models and going over the two versions that are available for network anomaly detection. Ultimately, the study offers an empirical assessment of the two types of models, just considering how well they detect and forecast anomalies overall. The results show that the unmarried-node anomaly detection performance of the model is superior. Simultaneously, the Adaptive Learning version is particularly effective in identifying anomalies among a few nodes. The fundamental reasons for the differences in the two fashions' overall performance are also examined in this research. This work provides a comparative analysis of two widely utilized algorithmic approaches: vector autoregressive models and community anomaly detection and prediction. Each method's effectiveness is assessed using two different network datasets: one based on real-world global measurements of latency and mobility ranges, and the other focused on a fictional community. The study also examines the trade-offs between employing the versus other modern and classic techniques, Markov Chain Monte Carlo, and Artificial Neural Networks for network anomaly detection. Finally, it provides an overview of the advantages and disadvantages of each technique as well as suggestions for improving performance.

show abstract