This paper deals with cooperative output tracking problems for heterogeneous networks of linear agents. To refine high-precision tracking performances of agents, a graph-based distributed learning control (DLC) law is proposed, for which a new bounded-initialization, bounded-updating (BIBU) stability property is explored under any bounded initial conditions. Moreover, a class of heterogeneous-to-homogeneous transformation methods is introduced, together with presenting feasible gain design conditions, for DLC. It is shown that with the designed DLC law, not only can the effect of the agents’ heterogeneous dynamics in performing DLC be well overcome, but also the BIBU stability and the robust cooperative output tracking of agents can be simultaneously accomplished. A simulation test is also implemented to verify the validity of our developed DLC results for heterogeneous vehicle networks.

中文翻译：

---

异构线性多智能体网络的分布式学习控制


本文研究线性代理异构网络的协作输出跟踪问题。为了提高智能体的高精度跟踪性能，提出了一种基于图的分布式学习控制（DLC）律，并在任何有界初始条件下探索了新的有界初始化、有界更新（BIBU）稳定性属性。此外，还介绍了一类 DLC 的异质到同质变换方法，并给出了可行的增益设计条件。结果表明，设计的DLC律不仅可以很好地克服智能体异质动态对DLC执行的影响，而且可以同时实现智能体的BIBU稳定性和鲁棒协作输出跟踪。还进行了模拟测试，以验证我们开发的异构车辆网络 DLC 结果的有效性。