Resilient Autonomous Control of Distributed Multiagent Systems in Contested Environments

Abstract

An autonomous and resilient controller is proposed for leader-follower multiagent systems under uncertainties and cyber-physical attacks. The leader is assumed nonautonomous with a nonzero control input, which allows changing the team behavior or mission in response to the environmental changes. A resilient learning-based control protocol is presented to find optimal solutions to the synchronization problem in the presence of attacks and system dynamic uncertainties. An observer-based distributed H∞ controller is first designed to prevent propagating the effects of attacks on sensors and actuators throughout the network, as well as to attenuate the effect of these attacks on the compromised agent itself. Nonhomogeneous game algebraic Riccati equations are derived to solve the H∞ optimal synchronization problem and off-policy reinforcement learning (RL) is utilized to learn their solution without requiring any knowledge of the agent's dynamics. A trust-confidence-based distributed control protocol is then proposed to mitigate attacks that hijack the entire node and attacks on communication links. A confidence value is defined for each agent based solely on its local evidence. The proposed resilient RL algorithm employs the confidence value of each agent to indicate the trustworthiness of its own information and broadcast it to its neighbors to put weights on the data they receive from it during and after learning. If the confidence value of an agent is low, it employs a trust mechanism to identify compromised agents and remove the data it receives from them from the learning process. The simulation results are provided to show the effectiveness of the proposed approach.

Resilient Consensus for Expressed and Private Opinions

“…A set of local filtering algorithms is put forward in [22], [23] to alleviate the influence of misbehaving agents, where each normal agent in the network discards the extremal values as compared to its own value. Resilient learning-based protocols are introduced in [24] to find optimal solutions to consensus problems in the presence of malicious attackers and uncertainties in system. By utilizing mobile detectors, it is shown in [25] that consensus can be maintained against Byzantine agents whose number is not restrained by the network connectedness.…”

Towards Security Joint Trust and Game Theory for Maximizing Utility: Challenges and Countermeasures

“…In addition to the explicit type of game model, researchers have established some game models based on different application scenarios [70][71][72][73][74]. Most of these game models are established with the following objectives as the ultimate goal: To establish trust relationships and increase the credibility of data, to defend against malicious attacks in the network, to help the system by choosing a better strategy, and to maximize network utility by achieving these goals [75][76][77][78][79][80][81].…”

“…Wu et al [60] Ruan et al [61] John et al [62] Wu et al [63] Jin and van Dijk [64] Yu et al [66] Abbass et al [68] Chica et al [69] Hu et al [6] Zhu et al [70] Ding et al [71] Zhang et al [72] Aloqaily et al [74] Rishwaraj et al [75] Pouryazdan et al [76] Chen et al [77] Lorenzo et al [78] Li et al [79] Aroyo et al [80] Moghadam and Modares [81] Cui et al [28] Yin et al [26] Chen et al [27] Li et al [14] According to whether the players cooperate, the game is divided into cooperative game and non-cooperative game. The common goal of both games can maximize network utility and improve network security.…”

“…The introduction of the game model improves the security of the network and against external and internal attacks [14,70,72,74,77]. There are also some game models that provide great help for network or system strategy selection [27,71,78,79,81], or addressed the problem of resource allocation [26]. In [28], the introduction of game theory optimizes the probability of successful transmission probabilities of data and the caching probabilities in some networks.…”

Consensus of Hybrid Multi-Agent Systems With Malicious Nodes

“…All results mentioned above concern consensus without fault tolerance ability and all agents in the network are assumed to be cooperative. However, in the real world, cyber physical attacks and malicious agents are not uncommon, which make the system vulnerable and undermine the consensus behavior [15], [16]. In this paper, we aim to study resilient consensus against malicious agents in hybrid multi-agent systems modeled by directed networks.…”