Software Implementation of the Neural Network Controller in Systems with Complementary Correction of the Control Action

The paper deals with the issues of software implementation of the neural network controller or the neural network in automatic control systems with complementary correction of the control action. The requirements to the neural network in the complementary correction circuit are determined. The algorithm of error back propagation with regularization is adopted as a training algorithm. Its software implementation in MATLAB Simulink package is given. The results of simulation of ACS with the neural network controller are presented. These control systems structures can be effectively used by various objects and technological processes.