Improved Passivity Analysis for Uncertain Neural Networks with Time-varying Delay

At the beginning of this paper, for neural network systems, the passivity of neural networks is studied. Then, the robust passivity of the uncertain neural networks with time-delay is analyzed. Based on the passive network theory, it can be obtained optimal passive results of the neural networks. The research shows that the less conservative passive condition of the neural networks can be obtained from constructing a new Lyapunov-Krasovskii functional and simplifying the quadratic terms of the new Lyapunov-Krasovskii functional's reciprocal. Meanwhile, we obtain the passivity condition which does not require all the symmetric matrices involved in the employed quadratic Lyapunov-Krasovskii functional's matrices to be positive definite. An example shows the results in this paper is effective.