Bumpless Transfer Control for Synchronization of Switched Neutral‐Type Neural Networks With a Reachable Set Strategy

• Published in: International journal of robust and nonlinear control Vol. 35; no. 6; pp. 2310 - 2323
• Main Authors: Li, Fang, Sang, Hong, Wang, Peng, Zhao, Ying, Ma, Yajing, Dimirovski, Georgi M.
• Format: Journal Article
• Language: English
• Published: Hoboken, USA John Wiley & Sons, Inc 01.04.2025; Wiley Subscription Services, Inc
• Subjects: bumpless transfer performance; Criteria; improved combined switching strategy; Neural networks; reachable set estimation; Simulation; switched neutral‐type neural networks; Synchronism; synchronization control
• ISSN: 1049-8923; 1099-1239
• DOI: 10.1002/rnc.7802

ABSTRACT This investigation primarily centers on the reachable‐set‐based bumpless transfer control (BTC) for the synchronization of switched neutral‐type neural networks (SNNNs). In order to mitigate the conservatism inherent in the traditional state‐dependent switching strategies (SDSSs) and combined switching strategies (CSSs), an improved CSS leveraging the historical information of neuron states and neutral delay is developed. By constructing a time‐dependent multiple Lyapunov‐Krasovskii functional (TDMLF) technique, a less conservative criterion for reachable set estimation (RSE) is first established. In the subsequent, the established design framework is further employed by the BTC for the synchronization of SNNNs. The corresponding synchronization criterion is derived, which ensures that the resultant synchronization error influenced by bounded external inputs can be confined to an anticipated bounded set. Also, the underlying control bumps at switching instants during switching instants are effectively constrained to a specific level. Ultimately, the practicability and superiority of the proposed design framework are confirmed via two simulation examples.