Shi Yongxia, Hu Qinglei, Shao Xiaodong, Shi Yang
IEEE Trans Neural Netw Learn Syst. 2023 Nov;34(11):8791-8801. doi: 10.1109/TNNLS.2022.3153077. Epub 2023 Oct 27.
This article addresses the event-triggered coordinated control problem for multiple Euler-Lagrange systems subject to parameter uncertainties and external disturbances. Based on the event-triggered technique, a distributed coordinated control scheme is first proposed, where the neural network-based estimation method is incorporated to compensate for parameter uncertainties. Then, an input-based continuous event-triggered (CET) mechanism is developed to schedule the triggering instants, which ensures that the control command is activated only when some specific events occur. After that, by analyzing the possible finite-time escape behavior of the triggering function, the real-time data sampling and event monitoring requirement in the CET strategy is tactfully ruled out, and the CET policy is further transformed into a periodic event-triggered (PET) one. In doing so, each agent only needs to monitor the triggering function at the preset periodic sampling instants, and accordingly, frequent control updating is further relieved. Besides, a parameter selection criterion is provided to specify the relationship between the control performance and the sampling period. Finally, a numerical example of attitude synchronization for multiple satellites is performed to show the effectiveness and superiority of the proposed coordinated control scheme.
本文研究了受参数不确定性和外部干扰影响的多个欧拉 - 拉格朗日系统的事件触发协同控制问题。基于事件触发技术，首先提出了一种分布式协同控制方案，其中采用基于神经网络的估计方法来补偿参数不确定性。然后，开发了一种基于输入的连续事件触发（CET）机制来调度触发时刻，这确保了控制命令仅在某些特定事件发生时才被激活。之后，通过分析触发函数可能的有限时间逃逸行为，巧妙地排除了CET策略中的实时数据采样和事件监测要求，并将CET策略进一步转换为周期性事件触发（PET）策略。这样一来，每个智能体只需要在预设的周期性采样时刻监测触发函数，从而进一步减轻了频繁的控制更新。此外，还提供了一个参数选择准则来指定控制性能与采样周期之间的关系。最后，进行了多颗卫星姿态同步的数值示例，以展示所提出的协同控制方案的有效性和优越性。
