Automated Adversary Emulation for Cyber-Physical Systems via Reinforcement Learning

被引:9
作者
Bhattacharya, Arnab [1 ]
Ramachandran, Thiagarajan [1 ]
Banik, Sandeep [3 ]
Dowling, Chase P. [2 ]
Bopardikar, Shaunak D. [3 ]
机构
[1] Pacific Northwest Natl Lab, Optimizat & Control Grp, Richland, WA 99352 USA
[2] Pacific Northwest Natl Lab, Informat Modeling & Analyt Grp, Richland, WA 99352 USA
[3] Michigan State Univ, Elect & Comp Engn, E Lansing, MI 48824 USA
来源
2020 IEEE INTERNATIONAL CONFERENCE ON INTELLIGENCE AND SECURITY INFORMATICS (ISI) | 2020年
关键词
Adversary Emulation; Reinforcement Learning; Cyber-Physical Security; Hybrid Attack Graph;
D O I
10.1109/isi49825.2020.9280521
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Adversary emulation is an offensive exercise that provides a comprehensive assessment of a system's resilience against cyber attacks. However, adversary emulation is typically a manual process, making it costly and hard to deploy in cyber-physical systems (CPS) with complex dynamics, vulnerabilities, and operational uncertainties. In this paper, we develop an automated, domain-aware approach to adversary emulation for CPS. We formulate a Markov Decision Process (MDP) model to determine an optimal attack sequence over a hybrid attack graph with cyber (discrete) and physical (continuous) components and related physical dynamics. We apply model-based and model-free reinforcement learning (RL) methods to solve the discrete-continuous MDP in a tractable fashion. As a baseline, we also develop a greedy attack algorithm and compare it with the RL procedures. We summarize our findings through a numerical study on sensor deception attacks in buildings to compare the performance and solution quality of the proposed algorithms.
引用
收藏
页码:1 / 6
页数:6
相关论文
共 16 条
  • [1] Angle Matthew G., 2019, IEEE Power and Energy Technology Systems Journal, V6, P172, DOI 10.1109/JPETS.2019.2923970
  • [2] Intelligent, Automated Red Team Emulation
    Applebaum, Andy
    Miller, Doug
    Strom, Blake
    Korban, Chris
    Wolf, Ross
    [J]. 32ND ANNUAL COMPUTER SECURITY APPLICATIONS CONFERENCE (ACSAC 2016), 2016, : 363 - 373
  • [3] Optimal Attack Strategies Subject to Detection Constraints Against Cyber-Physical Systems
    Chen, Yuan
    Kar, Soummya
    Moura, Jose M. F.
    [J]. IEEE TRANSACTIONS ON CONTROL OF NETWORK SYSTEMS, 2018, 5 (03): : 1157 - 1168
  • [4] Chhetri SR, 2019, DES AUT TEST EUROPE, P770, DOI [10.23919/date.2019.8715283, 10.23919/DATE.2019.8715283]
  • [5] Online Learning for Commercial Buildings
    Dong, Jin
    Ramachandran, Thiagarajan
    Im, Piljae
    Huang, Sen
    Chandan, Vikas
    Vrabie, Draguna L.
    Kuruganti, Teja
    [J]. E-ENERGY'19: PROCEEDINGS OF THE 10TH ACM INTERNATIONAL CONFERENCE ON FUTURE ENERGY SYSTEMS, 2019, : 522 - 530
  • [6] Dynamic Games in Cyber-Physical Security: An Overview
    Etesami, S. Rasoul
    Basar, Tamer
    [J]. DYNAMIC GAMES AND APPLICATIONS, 2019, 9 (04) : 884 - 913
  • [7] Hota A. R., 2018, Game Theory for Security and Risk Management, P157
  • [8] Automatic Hybrid Attack Graph (AHAG) Generation for Complex Engineering Systems
    Ibrahim, Mariam
    Alsheikh, Ahmad
    [J]. PROCESSES, 2019, 7 (11)
  • [9] Miehling E., 2015, P 2 ACM WORKSH MOV T, P67
  • [10] Miller D., 2018, TECHNICAL PAPER