Variable-sampling-period dependent global stabilization of delayed memristive neural networks based on refined switching
- PDF / 418,663 Bytes
- 16 Pages / 595 x 842 pts (A4) Page_size
- 17 Downloads / 166 Views
. RESEARCH PAPER .
November 2020, Vol. 63 212201:1–212201:16 https://doi.org/10.1007/s11432-019-2664-7
Variable-sampling-period dependent global stabilization of delayed memristive neural networks based on refined switching event-triggered control Zhilian YAN1 , Xia HUANG1* & Jinde CAO2 1
College of Electrical Engineering and Automation, Shandong University of Science and Technology, Qingdao 266590, China; 2 School of Mathematics, Southeast University, Nanjing 210096, China Received 11 June 2019/Accepted 2 September 2019/Published online 9 October 2020
Abstract This paper studies the stabilization problem of delayed memristive neural networks under eventtriggered control. A refined switching event-trigger scheme that switches between variable sampling and continuous event-trigger can be designed by introducing an exponential decay term into the threshold function. Compared with the existing mechanisms, the proposed scheme can enlarge the interval between two successively triggered events and therefore can reduce the amount of triggering times. By constructing a time-dependent and piecewise-defined Lyapunov functional, a less-conservative criterion can be derived to ensure global stability of the closed-loop system. Based on matrix decomposition, equivalent conditions in linear matrix inequalities form of the above stability criterion can be established for the co-design of both the trigger matrix and the feedback gain. A numerical example is provided to demonstrate the effectiveness of the theoretical analysis and the advantages of the refined switching event-trigger scheme. Keywords event-triggered control, delayed memristive neural networks, global stabilization, timedependent Lyapunov functional, variable sampling Citation Yan Z L, Huang X, Cao J D. Variable-sampling-period dependent global stabilization of delayed memristive neural networks based on refined switching event-triggered control. Sci China Inf Sci, 2020, 63(11): 212201, https://doi.org/10.1007/s11432-019-2664-7
1
Introduction
Over the past decade, memristive neural networks (MNNs), whose connection weights are implemented and determined through the use of memristors [1], have received considerable research attention [2–9]. Unlike the traditional neural networks, MNNs have more powerful storage and computing capacities, thus, resulting to their applicability in various application areas such as data storage, secure communication, and image processing [10,11]. It is a well-known fact that the stability of MNNs is the basic precondition for these applications. However, an MNN is naturally a switched system, and the switching law is majorly dependent on the evolution of system states. To be more specific, the MNN with n neurons consists of 2n subsystems. Such a switching characteristic may lead to undesirable dynamical phenomena such as oscillation, bifurcation, and even chaos. Furthermore, time-delays are inevitable in the hardware implementation of MNNs because of the finite switching speed of amplifiers [12]. They are often one of the primary
Data Loading...
