This work is concerned with mean-square bipartite containment of second-order multileader multiagent systems (MASs), contaminating compound noise, and antagonistic information under a fixed or Markovian switching signed topology. A new class of bipartite containment control protocols based on absolute velocity and relative position information is designed using signed graphs, and a time-varying control gain is introduced to eliminate the combined effect of additive and multiplicative noises. For the case under fixed topology, sufficient conditions for achieving bipartite containment are derived by using the Lyapunov function method. Then, it is extended to the case of nonlinear dynamics. Specific convergence values for all leaders and followers are obtained. For the case under Markovian switching topology, the boundedness of the agents' states and noise intensity is proved by employing the extended second-moment method. By utilizing properties of stochastic matrices and the ergodicity of Markov chains, the second moments of the bipartite containment errors are estimated to obtain mean-square bipartite containment conditions. In addition, the corresponding convergence rates of mean-square errors are explicitly expressed. The effectiveness of the theoretical results is finally verified through numerical simulations.
Zhang et al. (Tue,) studied this question.