Robust Data-Driven Compensation Iterative Learning Consensus Tracking Control for MIMO Multiagent Systems With Random Packet Dropouts
针对未知异构多输入多输出多智能体系统在随机丢包下的一致性跟踪问题，提出一种仅利用局部历史测量数据的分布式数据驱动补偿迭代学习控制方法，并扩展到有外部干扰和迭代变化拓扑的情形，理论分析和仿真验证了方法的稳定性和收敛性。
This article studies the consensus tracking control problem for unknown heterogeneous multi-input multi-output (MIMO) multiagent systems (MASs) with random packet dropouts. First, to reduce the impact of the packet dropouts, a distributed data-driven compensation iterative learning control (DDCILC) consensus tracking method is proposed, which utilizes only local historical measurements and employs a novel compensation technique for the MASs under random packet dropouts. Next, the DDCILC method is extended to controlling the MASs with external unmeasurable disturbances and iteration-varying topologies. The stability and convergence of the proposed approaches are rigorously analyzed under reasonable conditions. Compared with the existing results, the approaches proposed relax the requirement for the MASs graph structure, improve the iteration convergence speed and mitigate the deterioration of system control performance caused by packet dropouts. Finally, simulations are provided to verify the effectiveness of the proposed algorithms.