Neural Network-Based Optimizing Controller With On-Line Adaptation

2 Neural Network-Based Optimizing Controller With On-Line Adaptation 

In this section, a neural network-based design methodology is developed that utilizes the adaptability of neural networks to compensate for the time varying dynamical properties of smart structures. This formulation is designed to be implemented using the ETANN chip and also allows the designer to directly incorporate all the a priori information about the system that may be available. An important feature of this formulation is that it relies only on the experimental input/output data of the system for the design. The ability of neural networks to map nonlinear systems allows this formulation to be extended to incorporate nonlinearity in structural systems. 

A functional block diagram of the controller is shown in Fig. 4.13, where the structural system can be represented by 

The neural network in the controller block diagram has a model IV architecture with one hidden layer, as shown in Fig. 4.14. It is pretrained to the dynamics of the smart structural system using experimental input/output data. As shown in Fig. 4.15, the input vector to the network consists of n - - 1 samples of the plant input and m - - 1 samples of the plant output. The hidden and output layers have P and 1 neurons respectively. 

The activation function of the neurons in the hidden layer is the adaptive activation function (4.1). Models II and III are alternative neural network architectm es that can be used to model a d3oiamical system. Model III is similar to model IV except for the additional external adder and separate network for the plant input and output parts. Model III can be used to implement high-order dynamical swtem models using hardware neural networks like the ETANN. 

---