Passivity Analysis for Switched Generalized Delayed Neural Networks
Hong Sang, Hong Shan Nie, Jun Hui Fu · IEEE Transactions on Circuits & Systems II Express Briefs · 2022
This brief concerns the passivity issue for a class of continuous-time switched generalized neural networks (CSGNNs) with time-varying delay. Note that the existing related results about switched delayed neural networks are based on the time-dependent or the state-dependent switching strategies, where the time-dependent versions require the stability of each subnetwork, while the state-dependent versions can not avoid the frequent switching, chattering and Zeno problems. For handling this situation, the combined switching strategy is introduced on the basis of the constructed time-dependent multiple Lyapunov-Krasovskii functional approach. A new passivity criterion with less conservativeness is formulated, under which the addressed neural networks are passive with a guaranteed switching frequency. Also, each individual subnetwork can be non-passive. Finally, the employed simulation example substantiates the accuracy and superiority of the derived results.