Global exponential synchronization of discrete-time high-order switched neural networks and its application to multi-channel audio encryption

State Estimation for a Class of Discrete-Time BAM Neural Networks With Multiple Time-Varying Delays

“…The method directly uses the generalized matrix inverses and the definitions of GES, and it avoids the construction of any LKF; (2) The obtained sufficient conditions are composed of linear scalars inequalities that is easy to solve; (3) It is suitable for the more general neural network models after a small modification. For example, memristor-based NNs [40], inertial neural works [41] and high-order NNs [4], [42], [43].…”

“…I N recent years, many excellent results on neural networks (NNs) have been addressed extensively, since they were applicable to many fields such as pattern recognition, artificial intelligence, optimization, etc [1]- [4]. Generally, many NNs including biological NNs are composed of many interconnected man-made or/and natural dynamical units.…”

Optimized leader‐follower consensus control using combination of reinforcement learning and sliding mode mechanism for multiple robot manipulator system

“…Neural network (NN) has been demonstrated to be a universal approximator, and can be executed for the continuous and discrete nonlinear system controls. 28,29 Given a continuous nonlinear function F(𝜐) ∶ R n → R defined on a compact set Ω, its NN approximation can use the expression below…”

“…Neural network (NN) has been demonstrated to be a universal approximator, and can be executed for the continuous and discrete nonlinear system controls 28,29 . Given a continuous nonlinear function $$$ F\left(\upsilon \right):{R}^n\to R $$$ defined on a compact set $$$ \Omega $$$, its NN approximation can use the expression below $$$$ {F}_{NN}\left(\upsilon \right)={\Psi}^T\varphi \left(\upsilon \right), $$$$ where $$$ \Psi \in {R}^p $$$ and $$$ \varphi \left(\upsilon \right)={\left[{\varphi}_1\left(\upsilon \right),\dots, {\varphi}_p\left(\upsilon \right)\right]}^T\in {R}^p $$$ are the NN weight and basis function vector, respectively, $$$ {\varphi}_i\left(\upsilon \right) $$$ is determined by Gaussian function, that is, $$$ {\varphi}_i\left(\upsilon \right)=\mathit{\exp}\left[-{\left(\upsilon -{o}_i\right)}^T\left(\upsilon -{o}_i\right)/2{\varsigma}_i^2\right] $$$, on which …”

The Audio Encryption Approach uses a Unimodular Matrix and a Logistic Function

“…From the literature study, it appears that audio encryption using the unimodular matrix and logistic function has not been studied much. Therefore, this research is expected to provide contributions and solutions in the development of more secure and effective audio encryption [15], [16].…”