Time-Delay Nonlinear System Modelling via Delayed Neural Networks

In this paper, nonlinear systems on-line identification via delayed dynamic neural networks is studied. Dynamic series-parallel neural network model with time delay is persented and the stability conditions are derived using Lyapunov-Krasovskii approach. The conditions for passivity, asymptotic stability stability are established in some senses. All the results are described by linear matrix inequality (LMI). We conclude that the gradient algoritm for weight adjusment is stable and robust to any bounded uncertainties.