A Deep Reinforcement Learning-Based Secure Framework against Adversarial Attacks in Power System

Authors: Arshia Aflaki, Hadis Karimipour, Anik Islam

Abstract:

Generative Adversarial Attacks (GAAs) threaten critical sectors, ranging from fingerprint recognition to industrial control systems. Existing Deep Learning (DL) algorithms are not robust enough against this kind of cyber-attack. As one of the most critical industries in the world, the power grid is not an exception. In this study, a Deep Reinforcement Learning-based (DRL) framework assisting the DL model to improve the robustness of the model against GAAs is proposed. Real-world smart grid stability data, as an IIoT dataset, test our method and improve the classification accuracy of a DL model from around 57% to 96%.

Keywords: Generative Adversarial Attack, Deep Reinforcement Learning, deep learning, IIoT, Generative Adversarial Networks, power system.

References:

[1] Arturo S Bretas, Newton G Bretas, and Breno EB Carvalho. Further contributions to smart grids cyber-physical security as a malicious data attack: Proof and properties of the parameter error spreading out to the measurements and a relaxed correction model. International Journal of Electrical Power & Energy Systems, 104:43–51, 2019.
[2] Tanveer Ahmad, Dongdong Zhang, Chao Huang, Hongcai Zhang, Ningyi Dai, Yonghua Song, and Huanxin Chen. Artificial intelligence in sustainable energy industry: Status quo, challenges and opportunities, 2021.
[3] Ian J. Goodfellow, Jean Pouget-Abadie, Bing Xu Mehdi Mirza, David Warde-Farley, Sherjil Ozairy, Aaron Courville, and Yoshua Bengioz. Generative adverserial nets. Universite de Montr ´ eal, 2014.
[4] Qianwei Zhou, Margarita Zuley, Yuan Guo, Lu Yang, Bronwyn Nair, Adrienne Vargo, Suzanne Ghannam, Dooman Arefan, and Shandong Wu. A machine and human reader study on ai diagnosis model safety under attacks of adversarial images. Nature Communications, 12, 2021.
[5] Xinyu Wang, Xiaoyuan Luo, Yuyan Zhang, and Xinping Guan. Detection and isolation of false data injection attacks in smart grids via nonlinear interval observer. IEEE Internet of Things Journal, 6, 2019.
[6] Zhe Wang, Wei Cheng, and Chunlin Li. Dos attack detection model of smart grid based on machine learning method. 2020.
[7] Mohsen Ghafouri, Minh Au, Marthe Kassouf, Mourad Debbabi, Chadi Assi, and Jun Yan. Detection and mitigation of cyber attacks on voltage stability monitoring of smart grids. IEEE Transactions on Smart Grid, 11, 2020.
[8] Arshia Aflaki, Mohsen Gitizadeh, and Burak Kantarci. Accuracy improvement of electrical load forecasting against new cyber-attack architectures. Sustainable Cities and Society, 77, 2022.
[9] Alireza Pourdaryaei, Amidaddin Shahriari, Mohammad Mohammadi, Mohammad Reza Aghamohammadi, Mazaher Karimi, and Kimmo Kauhaniemi. A new approach for long-term stability estimation based on voltage profile assessment for a power grid. Energies, 16(5), 2023.
[10] Lu Yang, Qing Song, and Yingqi Wu. Attacks on state-of-the-art face recognition using attentional adversarial attack generative network. Multimedia Tools and Applications, 80, 2021.
[11] Volodymyr Mnih, Koray Kavukcuoglu, David Silver, Andrei A. Rusu, Joel Veness, Marc G. Bellemare, Alex Graves, Martin Riedmiller, Andreas K. Fidjeland, Georg Ostrovski, Stig Petersen, Charles Beattie, Amir Sadik, Ioannis Antonoglou, Helen King, Dharshan Kumaran, Daan Wierstra, Shane Legg, and Demis Hassabis. Human-level control through deep reinforcement learning. Nature, 518, 2015.
[12] John A. Bland, Mikel D. Petty, Tymaine S. Whitaker, Katia P. Maxwell, and Walter Alan Cantrell. Machine learning cyberattack and defense strategies. Computers and Security, 92, 2020.
[13] Rajesh Ganesan, Sushil Jajodia, Ankit Shah, and Hasan Cam. Dynamic scheduling of cybersecurity analysts for minimizing risk using reinforcement learning. ACM Transactions on Intelligent Systems and Technology, 8, 2016.
[14] Lirong Liu, Hao Ma, Siyun Chen, Xinming Tang, Junfeng Xie, Gang Huang, and Fan Mo. Image-translation-based road marking extraction from mobile laser point clouds. IEEE Access, 8, 2020.
[15] Vadim Arzamasov, Klemens Bohm, and Patrick Jochem. Towards concise models of grid stability. 2018.
[16] Deep q-learning — an introduction to deep reinforcement learning.
[17] Zahra Moti, Sattar Hashemi, Hadis Karimipour, Ali Dehghantanha, Amir Namavar Jahromi, Lida Abdi, and Fatemeh Alavi. Generative adversarial network to detect unseen internet of things malware. Ad Hoc Networks, 122, 2021.
[18] Rahim Taheri, Mohammad Shojafar, Mamoun Alazab, and Rahim Tafazolli. Fed-iiot: A robust federated malware detection architecture in industrial iot. IEEE Transactions on Industrial Informatics, 2020.
[19] Mohammad Mehedi Hassan, Md Rafiul Hassan, Shamsul Huda, and Victor Hugo C. De Albuquerque. A robust deep-learning-enabled trust-boundary protection for adversarial industrial iot environment. IEEE Internet of Things Journal, 8, 2021.
[20] sklearn.preprocessing.minmaxscaler — scikit-learn 1.0.2 documentation.