In this paper, the D-type iterative learning control (ILC) protocol based on the local neighbor information is designed to achieve tracking synchronization for linearly coupled reaction-diffusion neural networks in presence of time delay and iteration-varying switching topology under a repetitive environment. Firstly, based on non-collocated sensors and actuators network, the proposed D-type ILC update law can realize tracking synchronization by utilizing output tracking errors. Then, by virtue of the contraction mapping principle, the sufficient convergence conditions of tracking synchronization errors are presented under the fixed commutation topology. Subsequently, the synchronization conclusions are extended to the iteration-varying commutation topology scenario. Finally, two numerical examples are provided to verify the efficacy of the obtained results.

本文设计了一种基于局部邻居信息的D型迭代学习控制（ILC）协议，以在重复环境下存在时滞和迭代变切换拓扑的情况下实现线性耦合反应扩散神经网络的跟踪同步。 。首先，基于非并置的传感器和执行器网络，提出的D型ILC更新定律可以利用输出跟踪误差实现跟踪同步。然后，根据压缩映射原理，给出了固定换相拓扑下跟踪同步误差的充分收敛条件。随后，将同步结论扩展到迭代可变换向拓扑方案。最后，