Distributed Bipartite Consensus for Multi-Agent Systems Via Data-Driven Sliding Mode Scheme

This paper investigates a fully distributed bipartite consensus tracking problem for nonlinear discrete-time Multi-agent systems (MASs) with unknown dynamics and antagonistic interactions. A fully distributed data-driven sliding mode bipartite consensus (DSMBC) approach is proposed. The convergence of the proposed method is no longer related to the format of the reference trajectory, including time-varying and time-invariant trajectories. Moreover, the strongly connected requirement is no longer needed. Firstly, a bipartite combined measurement error function is formulated to transfer the bipartite consensus issue into the consensus issue. Then, an enhanced compact form dynamic linearization data mode is established by employing the input/output data of the MASs. After that, the DSMBC is constructed, and the proposed algorithm's convergence is proved, showing that each agent's bipartite consensus tracking error is cut to a small region around the origin. Finally, two examples are presented, and the results further demonstrate the correctness and effectiveness of the proposed scheme, where the MASs can tackle both time-varying and time-invariant tracking tasks.