Passivity Analysis of Neural Networks with Time Delay

CG Li,XF Liao
DOI: https://doi.org/10.1109/tcsii.2005.849023
2005-01-01
IEEE Transactions on Circuits and Systems II Analog and Digital Signal Processing
Abstract:The passivity conditions for delayed neural networks (DNNs) are considered in this paper. We firstly derive the passivity condition for DNNs without uncertainties, and then extend the result to the case with time-varying parametric uncertainties. The proposed approach is based on a Lyapunov-Krasovskii functional construction. The passivity conditions are presented in terms of linear matrix inequalities, which can be easily solved by using the effective interior-point algorithm. Numerical examples are also given to demonstrate the effectiveness of the theoretical results.
What problem does this paper attempt to address?