TABLE OF CONTENTS Different control schemes were already introduced in Section 1.6, in the context of fuzzy systems. In fact, as fuzzy systems and neural networks are both universal approximators, they are used in the same control structures. Therefore, in this section, rather than introducing them again, the reader is referred to the corresponding section in Chapter 1, and important references, in terms of neural networks, are given. Good surveys on the application of neural networks for control can be found in, for instance, [79] and [80]. Also, please see the papers in the special issue [81].
Gain scheduling control is covered in Section 1.6.1. Neural network applications of this approach can be found in, for instance, [82], where RBFs are used as a gain-scheduling controller for the lateral motion of a propulsion controlled aircraft, and in [83], where a regularised neural network is used to improve the performance of a classic continuous-parameter gain-scheduling controller.
The most straightforward application of neural networks is adaptive inverse control. This topic is covered in Section 1.6.3, and important neural network applications can be found in [84] [87].
Internal model control with fuzzy systems is introduced in Section 1.6.4. This approach was introduced, in the context of neural networks in [88]. This technique was used for the control of a bioreactor in [89], for pH neutralisation, using local model networks in [90]
