Standard neural network model for robust stabilization of recurrent neural networks

The paper applies Lyapunov stability theory and S-procedure technique to investigate the robust stabilization problem of standard neural network model(SNNM). State-feedback controllers are designed to guarantee the global asymptotical stability of SNNM with norm-bounded uncertainties. The control law presented are formulated as linear matrix inequalities to be easily solved. Most of the existing recurrent neural networks can be transformed into SNNMs to be synthesized in a unified way. An example shows the effectiveness of this method.