Robust Resilient H ∞ State Estimation for Time-varying Recurrent Neural Networks Subject to Probabilistic Quantization Under Variance Constraint

This paper is concerned with the robust resilient H ∞ state estimation problem for time-varying recurrent neural networks (TVRNNs) with probabilistic quantization under variance constraint. Here, a situation is considered where the signals are quantized before entering the network, and the occurrence probability is assumed to be known. In addition, during the design of the state estimation algorithm, the additive variation of the estimator gain is considered to reflect the parameter deviation that may occur during the execution. The main purpose is to design a finite-horizon resilient state estimation algorithm such that, in the presence of probabilistic quantization and estimator gain perturbation, some sufficient criteria are obtained for the estimation error system to satisfy the prescribed H ∞ performance requirement within the finite-horizon and the error variance boundedness. Finally, a numerical example is conducted to verify the feasibility of the presented estimation algorithm against the probabilistic quantization and estimator gain perturbation.