Particularities of Identification and Tuning of Linear Systems with State Controllers Using Artificial Neural Networks

At present, state controllers are used to improve the quality of mechatronic systems control. The high potential of systems with such controllers can be implemented only if automatic tuning means are available. Since tuning is performed in real time mode, which places higher demands on performance, to reduce tuning time it is proposed to use an artificial neural network. However, under interference in measuring channels, the quality of parameter identification in the controlled object is significantly lowered. In this regard, the purpose of this study is to find the optimal composition of measuring channels at the network input, making it possible to minimize the impact of interference on estimates of object parameters to improve control quality. In the present study, an algorithm is proposed to optimize the positioning of measuring channels in the task of identifying and tuning control systems. As an optimality criterion, we use the information signal-to-interference power ratio at network inputs.