Concluding Remarks and Further Research Directions

Abstract

Owing to their powerful properties, artificial neural networks have become a popular choice in terms of problems occurring in control theory. Furthermore, the following two trends in modern control: robust control and fault-tolerant control can be effectively realized using appropriate neural-network architecture. This monograph is devoted to the selected designs of the robust and fault-tolerant control systems for nonlinear processes. The reported approaches mainly use the capability of a neural network to learn from historical data and to approximate nonlinear functions with an assumed accuracy. These two properties are extremely useful when dealing with nonlinear industrial plants for which a mathematical model is unknown or is very expensive to determine.