ALGORITHMS FOR SYNTHESIS OF ADAPTIVE CONTROL SYSTEMS BASED ON THE NEURAL NETWORK APPROACH
Abstract
Here are algorithms for the synthesis of adaptive control systems based on the neural network approach presented. The issues of choice of architectures of neural networks that produce parametric identification of the object regulation. Adaptative neural networks that reproduce the optimal parameters of implemented controllers, methods of forming training samples for their training are considered. The other question of using the theory of artificial neural networks for approximating the functional of the controller's tuning parameters are determined. Structures of synthesized neural networks of parametric identification of objects with self-alignment in an open loop ACS are proposed. Algorithms made it possible to effectively solve the problems of adaptive systems synthesis based on the neural network approach in control systems of technological processes.