H∞ State Estimation for Discrete-Time Delayed Systems of the Neural Network Type With Multiple Missing Measurements

Area of Science:

• Control Systems Engineering
• Nonlinear System Analysis
• Stochastic Systems

Background:

• Discrete-time nonlinear systems, including neural networks, often suffer from random time-varying delays and missing measurements.
• Accurate state estimation is crucial for controlling these complex systems but is challenging under such uncertainties.

Purpose of the Study:

• To design an H∞ state estimator for discrete-time nonlinear systems with random time-varying delays and multiple missing measurements.
• To ensure global mean square stability of the estimation error dynamics.
• To develop less conservative estimation results by considering delay variations and probabilities.

Main Methods:

• Construction of a Luenberger-like estimator utilizing imperfect output data.
• Application of Lyapunov stability theory and stochastic methods to analyze H∞ performance.
• Deduction of H∞ estimator gains incorporating delay range and probability distributions.

Main Results:

• The proposed H∞ estimator guarantees global mean square stability for the augmented system.
• The method effectively handles random time-varying delays and multiple missing measurements.
• Incorporating delay characteristics leads to less conservative and more accurate state estimation.

Conclusions:

• The developed H∞ state estimation approach is effective for discrete-time nonlinear neural network-type systems with significant uncertainties.
• The proposed method provides a robust framework for state estimation in systems with random delays and data loss.
• Validation through examples confirms the practical applicability and improved performance of the proposed estimator.