Quasi-Consensus Control for Stochastic Multiagent Systems: When Energy Harvesting Constraints Meet Multimodal FDI Attacks

In this article, the quasi-consensus control problem is investigated for a class of stochastic nonlinear time-varying multiagent systems (MASs). The innovation points of this research can be highlighted as follows: first of all, the dynamics of the plant are stochastic, nonlinear, and time varying, which resembles the natural systems in practice closely. Meanwhile, an energy harvesting protocol is put forward to collect adequate energy from the external environment. Second, as a generalization of the existing result, the ultimate control objective is quasi-consensus in a probabilistic sense, that is, designing a distributed control protocol in order that the probability of centering the allowable region for the states of each agent is larger than some predetermined values. Third, the MASs are subject to false data-injection (FDI) attacks, and a more general multimodal FDI model is proposed. On the basis of the probabilistic-constrained analysis technique and the recursive linear matrix inequalities (RLMIs), sufficient conditions are provided to guarantee the probabilistic quasi-consensus property. To derive the controller gains, an optimal probabilistic-constrained algorithm is designed by solving a convex optimization problem. Finally, two examples are provided to substantiate the validity of the proposed framework.