A3D: Attention-based auto-encoder anomaly detector for false data injection attacks

Kundu, Arnav; Sahu, Abhijeet; Serpedin, Erchin; Davis, Katherine

doi:10.1016/j.epsr.2020.106795

Cited by 43 publications

(37 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…RESLab analyses the challenges of incorporating MiTM attacks under different polling rates and numbers of polled DNP3 outstations for a large-scale power system using queuing theory. Extensive research has been proposed on defence against FDI attacks [44,45]. However, works that address FDI attacks tend to make unrealistic assumptions on the adversary's capabilities in the communication network; RESLab remedies this by enabling its emulation of the communication system and highfidelity FDI attacks.…”

Section: Criteria For Design Decisionsmentioning

confidence: 99%

Design and evaluation of a cyber‐physical testbed for improving attack resilience of power systems

Sahu

Wlazlo

Mao

et al. 2021

IET Cyber-Phy Sys Theory & Ap

Self Cite

View full text Add to dashboard Cite

A power system is a complex cyber-physical system whose security is critical to its function. A major challenge is to model, analyse and visualise the communication backbone of the power systems concerning cyber threats. To achieve this, the design and evaluation of a cyber-physical power system (CPPS) testbed called Resilient Energy Systems Lab (RESLab) are presented to capture realistic cyber, physical, and protection system features. RESLab is architected to be a fundamental platform for studying and improving the resilience of complex CPPS to cyber threats. The cyber network is emulated using Common Open Research Emulator (CORE), which acts as a gateway for the physical and protection devices to communicate. The physical grid is simulated in the dynamic time frame using Power World Dynamic Studio (PWDS). The protection components are modelled with both PWDS and physical devices including the SEL Real-Time Automation Controller (RTAC). Distributed Network Protocol 3 (DNP3) is used to monitor and control the grid. Then, the design is exemplified and the tools are validated. This work presents four case studies on cyberattack and defence using RESLab, where we demonstrate false data and command injection using Man-in-the-Middle and Denial of Service attacks and validate them on a large-scale synthetic electric grid.This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

show abstract

Section: Criteria For Design Decisionsmentioning

confidence: 99%

Design and evaluation of a cyber‐physical testbed for improving attack resilience of power systems

Sahu

Wlazlo

Mao

et al. 2021

IET Cyber-Phy Sys Theory & Ap

Self Cite

View full text Add to dashboard Cite

show abstract

“…In recent years, to effectively detect FDIAs in AC-model systems, some deep learning based methods have been proposed [10]- [14]. Kundu et al [11] proposed an auto-encoder-based unsupervised learning method to detect FDIAs in AC-model systems. In addition, Zhang et al [14] proposed a semi-supervised deep learning approach by integrating an auto-encoder into a generative adversarial network.…”

Section: Introductionmentioning

confidence: 99%

“…In addition, Zhang et al [14] proposed a semi-supervised deep learning approach by integrating an auto-encoder into a generative adversarial network. Compared with the method in [11], labeled false measurement data are used to train the network model. In both methods, the measurement data are used as the training dataset.…”

Section: Introductionmentioning

confidence: 99%

“…Because most of those methods mainly focus on temporal relationships between time-series measurement data and pay less attention to the spatial relationship of the measurement data between buses and transmission lines. For example, the auto-encoderbased method proposed in [11] only considers the temporal relationship of normal time-series measurement data. Because the measurement data at a particular time step are treated as a one-dimensional input, the spatial relationship between buses and lines is not considered in the power grid.…”

Section: Introductionmentioning

confidence: 99%

“…Yu et al [13] proposed a gated recurrent unit (GRU) network based detector to model the temporal relationship between a time-series measurement data. Compared with the two aforementioned methods [11], [14], one major difference is that the estimated state values instead of the original measurement data are used as the training dataset to train the FDIA detector. However, the estimated state values obtained from measurement data cannot thoroughly represent the spatial-temporal data pattern of the measurement data.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

A Subgrid-Oriented Privacy-Preserving Microservice Framework Based on Deep Neural Network for False Data Injection Attack Detection in Smart Grids

Yin

Zhu

2022

IEEE Trans. Ind. Inf.

View full text Add to dashboard Cite

False data injection attacks (FDIAs) have recently become a major threat to smart grids. Most of the existing FDIAS detection methods have focused on modeling the temporal relationship of time-series measurement data, but have paid less attention to the spatial relationship between bus/line measurement data, and have failed to consider the relationship between sub-grids. To address these issues, we propose a sub-grid-oriented microservice framework by integrating a well-designed spatialtemporal neural network for FDIA detection in AC-model power systems. First, a well-designed neural network was developed to model the spatial-temporal relationship of bus/line measurements for sub-grids. A microservice-based supervising network is then proposed for integrating the representation features obtained from sub-grids for the collaborative detection of FDIAs. To evaluate the proposed framework, three types of FDIA datasets are generated based on a public benchmark power grid. Case studies on the FDIA datasets show that our method outperforms state-of-the-art methods for FDIA detection in these datasets.

show abstract

False Data Injection Attack Detection Using Machine Learning in Industrial Internet of Things

Hafizunisa,

Rai,

Sinha

2024

Optimized Computational Intelligence Driven Decision‐Making

View full text Add to dashboard Cite

A3D: Attention-based auto-encoder anomaly detector for false data injection attacks

Cited by 43 publications

References 20 publications

Design and evaluation of a cyber‐physical testbed for improving attack resilience of power systems

Design and evaluation of a cyber‐physical testbed for improving attack resilience of power systems

A Subgrid-Oriented Privacy-Preserving Microservice Framework Based on Deep Neural Network for False Data Injection Attack Detection in Smart Grids

False Data Injection Attack Detection Using Machine Learning in Industrial Internet of Things

Contact Info

Product

Resources

About