Non-fragile state observation for delayed memristive neural networks: Continuous-time case and discrete-time case

The topic of non-fragile observation for memristive neural networks with both continuous-time and discrete-time cases are provided in this paper. By endowing the Lyapunov technique, the corresponding sufficient criteria for the stability findings are furnished in the form of linear matrix inequalities (LMIs), of which, the desired observer gains can be calculated via the LMIs. What is the difference lies that the driven memristive neural networks are recast into models with interval parameters when considering the fact that the parameters of memrisitve model are state-dependent, which lead to parameter mismatch issue when different initial values are given. Thus, a new robust control method is introduced to tackle with the target model. Finally, the analytical design are substantiated with numerical results.