An overview on multi-agent consensus under adversarial attacks

Ishii, Hideaki; Wang, Yuan; Feng, Shuai

doi:10.1016/j.arcontrol.2022.01.004

Cited by 60 publications

(12 citation statements)

References 114 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The auxiliary controller plays a crucial role in determining a consensus coupling voltage and regulating the energy flow between the MGs and the main energy grid. To achieve these goals, optimization strategies are employed, focusing on power loss mitigation, energy efficiency, and maximizing economic advantages [58][59][60][61][62]. This auxiliary control system contributes significantly to the overall stability and optimal performance of MGs.…”

Section: Auxiliary Control In Mgsmentioning

confidence: 99%

The nexus of energy in microgrids: A review on communication barriers in distributed networks auxiliary controls

Ahmed,

Basit,

e Mustafa

et al. 2023

IET Generation Trans & Dist

View full text Add to dashboard Cite

Microgrids are a new concept resulting from decentralized power sources, which have changed energy utility grid topography. As this promising field is penetrated, control must be streamlined immediately. This paper analyses key control mechanisms in energy utility networks, including centralized, decentralized, and distributed setups. The focus is on the auxiliary category as the various benefits and drawbacks of each approach is investigated. Information interchange architectures that enable multiple control techniques in the auxiliary control domain are examined using distributed energy networks. The research emphasizes the necessity of auxiliary control in ensuring reliable and effective utility infrastructure operation but notes the lack of a comprehensive scholarly reference. Additionally, the study explores key concerns in utility grids' auxiliary control mechanisms, emphasizing information‐sharing system limits. The goal is to demonstrate the importance of auxiliary control and evaluate its challenges, such as latency, which require new solutions for a seamless transition to a sustainable energy future through rigorous academic assessment. At the auxiliary level, complex, decentralized regulatory mechanisms are recommended to handle transmission channel bandwidth and operational overhead in energy utility networks. These innovative measures have the potential to transform the fundamental structure of our energy landscape, heralding in unprecedented efficiency and sustainability.

show abstract

Section: Auxiliary Control In Mgsmentioning

confidence: 99%

The nexus of energy in microgrids: A review on communication barriers in distributed networks auxiliary controls

Ahmed,

Basit,

e Mustafa

et al. 2023

IET Generation Trans & Dist

View full text Add to dashboard Cite

show abstract

“…Furthermore, some restrictions on the communication network architecture are needed for certain types of cyberattacks. For instance, when a certain number of agents are attacking and being adversarial, with a specific resilient control scheme, such as mean-subsequence reduced (MSR) algorithms, sufficient conditions regarding the robustness of graphs are derived aiming to mitigate the negative impacts from the adversarial and resilient consensus [13]. Notably, MSR and double-integrator position based MSR (DP-MSR) algorithms were investigated in [69], where some necessary and sufficient conditions regarding the topology restrictions were given for resilient consensus of NASs with time-varying communication graphs and under scenarios with misbehaving agents.…”

Section: Recent Proactive Defense Strategies For Securing Complex Cyb...mentioning

confidence: 99%

“…Table 1 illustrates some recent surveys or overviews on the security issues in multiagent systems (MASs) [11][12][13], networked control systems [14,15], or CPSs [16]. The authors note that, at the time of writing, there is a scarcity of comprehensive surveys that examine the state-of-the-art advances and novel methodologies in secure control for complex cyberphysical networks.…”

Section: Introductionmentioning

confidence: 99%

A Brief Survey of Recent Advances and Methodologies for the Security Control of Complex Cyber–Physical Networks

Wan

Alsaedi

2023

Sensors

View full text Add to dashboard Cite

Complex cyber–physical networks combine the prominent features of complex networks and cyber–physical systems (CPSs), and the interconnections between the cyber layer and physical layer usually pose significant impacts on its normal operation. Many vital infrastructures, such as electrical power grids, can be effectively modeled as complex cyber–physical networks. Given the growing importance of complex cyber–physical networks, the issue of their cybersecurity has become a significant concern in both industry and academic fields. This survey is focused on some recent developments and methodologies for secure control of complex cyber–physical networks. Besides the single type of cyberattack, hybrid cyberattacks are also surveyed. The examination encompasses both cyber-only hybrid attacks and coordinated cyber–physical attacks that leverage the strengths of both physical and cyber attacks. Then, special focus will be paid to proactive secure control. Reviewing existing defense strategies from topology and control perspectives aims to proactively enhance security. The topological design allows the defender to resist potential attacks in advance, while the reconstruction process can aid in reasonable and practical recovery from unavoidable attacks. In addition, the defense can adopt active switching-based control and moving target defense strategies to reduce stealthiness, increase the cost of attacks, and limit the attack impacts. Finally, conclusions are drawn and some potential research topics are suggested.

show abstract

“…Adversarial attacks and defenses in communicative MARL receive much less attention when compared to their counterparts in reinforcement learning (Mu et al 2022). Current frameworks in this area (Ishii, Wang, and Feng 2022;Yuan et al 2023) commonly follow the principle of passive defense, where agents treat all received messages equally and try to make relative safe decisions. Since perturbed messages are mixed with useful messages, this indiscriminate reception may lead to the result of "garbage in, garbage out".…”

Section: Introductionmentioning

confidence: 99%

Robust Communicative Multi-Agent Reinforcement Learning with Active Defense

Yu,

Qiu,

Yao

et al. 2024

AAAI

View full text Add to dashboard Cite

Communication in multi-agent reinforcement learning (MARL) has been proven to effectively promote cooperation among agents recently. Since communication in real-world scenarios is vulnerable to noises and adversarial attacks, it is crucial to develop robust communicative MARL technique. However, existing research in this domain has predominantly focused on passive defense strategies, where agents receive all messages equally, making it hard to balance performance and robustness. We propose an active defense strategy, where agents automatically reduce the impact of potentially harmful messages on the final decision. There are two challenges to implement this strategy, that are defining unreliable messages and adjusting the unreliable messages' impact on the final decision properly. To address them, we design an Active Defense Multi-Agent Communication framework (ADMAC), which estimates the reliability of received messages and adjusts their impact on the final decision accordingly with the help of a decomposable decision structure. The superiority of ADMAC over existing methods is validated by experiments in three communication-critical tasks under four types of attacks.

show abstract

An overview on multi-agent consensus under adversarial attacks

Cited by 60 publications

References 114 publications

The nexus of energy in microgrids: A review on communication barriers in distributed networks auxiliary controls

The nexus of energy in microgrids: A review on communication barriers in distributed networks auxiliary controls

A Brief Survey of Recent Advances and Methodologies for the Security Control of Complex Cyber–Physical Networks

Robust Communicative Multi-Agent Reinforcement Learning with Active Defense

Contact Info

Product

Resources

About