The setting neuro-controller in the problem of control of dynamic plants
DOI:
https://doi.org/10.18372/2310-5461.23.7391Keywords:
dynamic plant, neuro-controller, algorithm control, error situation, quasi-optimal to timeAbstract
The paper considers the problem of setting neurocontroller taking part in the control of a dynamic object with unknown parameters. The control system consists of performance unit that have neurocontroller, measurement unit that consist of sensors movement, system setting and logic unit. Setting neurocontroller carried out by his learning as a result of the next test. Feature of the algorithm is to fix the faulty control situations that are introduced in the learning algorithm, giving it the properties of the self-learning, this data is input data for logic unit. The proposed algorithm provides a quasi-optimal on time and accuracy of control in the control of a dynamic object. A version of the technical implementation of the control system neurocontroller is proposed, as well as the results of its modeling.
References
Ashour M., Aboshosha A. Adaptive neural control of NPR // Proceeding of A1 Azhar Engineering Fifth International Conference AEIC-97, Cairo, Egypt, December 19-22, 1997.
Noriega J.R., A direct adaptive neural network control for unknown nonlinear system and its application / Noriega J.R., Wang H. // IEEE Trans. Neural networks. - 9. - №1. - 1998. - P.27-34.
Efe M.O. A novel error critic for variable structure control with an ADALINE / M.O. Efe // Transactions of the Institute of Measurement and Control 24,5. - 2002. - P. 403-415.
Пат. № 81997 Україна, МПК G 05 B 13/00, G 05 B 11/00. Спосіб настроювання вагових коефіцієнтів системи “Adaline” при керуванні динамічними об’єктами / Кучерів Д.П.; заявник та патентовласник ЦНДІ ОВТ ЗС України. - Заявл. 2.06.06; опубл. 25.02.08, Бюл. № 4. - 6 с.
Widrow B., Groner G.F., Hu M.J., Smith F.W., Specht D., Talbert L.R. Practical application for adaptive data processing systems. - WESCON Techn. Papers, No 7, 11/4, 1963.
