|
Article information
2020 , Volume 25, ¹ 3, p.111-118
Voevoda A.A., Romannikov D.O.
The synthesis method of regulators for multichannel systems using neural networks
The problem for synthesis of automatic control systems is hard, especially for multichannel objects. One of the approaches is the use of neural networks. For the approaches that are based on the use of reinforcement learning, there is an additional issue — supporting of range of values for the set points. The method of synthesis of automatic control systems using neural networks and the process of its learning with reinforcement learning that allows neural networks learning for supporting regulation is proposed in the predefined range of set points. The main steps of the method are 1) to form a neural net input as a state of the object and system set point; 2) to perform modelling of the system with a set of randomly generated set points from the desired range; 3) to perform a one-step of the learning using the Deterministic Policy Gradient method. The originality of the proposed method is that, in contrast to existing methods of using a neural network to synthesize a controller, the proposed method allows training a controller from an unstable initial state in a closed system and set of a range of set points. The method was applied to the problem of stabilizing the outputs of a two-channel object, for which stabilization both outputs and the first near the input set point is required
[full text] [link to elibrary.ru]
Keywords: neural network, control, regulator, multichannel system, closed system
doi: 10.25743/ICT.2020.25.3.012
Author(s):
Voevoda Aleksandr Aleksandrovich
Dr. , Professor
Position: Professor
Office: Novosibirsk State Technical University
Address: 630087, Russia, Novosibirsk, Karl Marx Ave. 20
E-mail: voevoda@ucit.ru
Romannikov Dmitry Olegovich
PhD. , Associate Professor
Position: Associate Professor
Office: Novosibirsk State Technical University
Address: 630087, Russia, Novosibirsk, Karl Marx Ave. 20
E-mail: dmitry.romannikov@gmail.com
References:
1. Krizhevsky A., Sutskever I., Hinton G.E. ImageNet classification with deep convolutional neuralnetworks. Proc. of the Neural Information Processing Systems, New York, USA, 2012. Association for Computing Machinery. 2012; 30(6):1097–1105. 2012: 1097–1105.
2. Graves A., Mohamed A., Hinton G.E. Speech recognition with deep recurrent neural networks. Proc.of the Intern. Conf. on Acoustics, Speech and Signal Processing (ICASSP), Vancouver, BC. 2013: 6645–6649. Available at: https://ieeexplore.ieee.org/document/6638947
3. Deng L., Hinton G. E., Kingsbury B. New types of deep neural network learning for speech recognitionand related applications: An overview. Proc. of the Intern. Conf. on Acoustics, Speech and Signal Processing (ICASSP), Vancouver, BC. 2013: 8599–8603. Available at: https://ieeexplore.ieee. org/document/6639344)
4. Voevoda A.A. Stabilizatsiya dvukhmassovoy sistemy: polinomial’nyy metod sinteza dvukhkanal’noysistemy [Stabilization of a two-mass system: a polynomial method for the synthesis of a two-channel system]. Sbornik nauchnykh trudov NGTU. 2010; 4(62):13–24. (In Russ.)
5. Sutton R., Barto A. Reinforcement learning: An introduction. Cambridge: MIT Press; 2018: 1328.
6. Mnih V., Kavukcuoglu K., Silver D., Graves A., Antonoglou I., Wierstra D., Riedmiller M. Playingatari with deep reinforcement learning. Available at: https://arxiv.org/abs/1312.5602 (accessed 27.05.200).
7. Hester T., Vecerik M., Pietquin O., Lanctot M., Schaul T., Piot B., Horgan D., Quan J., Sendonaris A.,Dulac-Arnold G., Osband I., Agapiou J., Leibo J.Z., Gruslys A. Deep Q-learning from demonstrations. Available at: https://arxiv.org/abs/1704.03732 (accessed 27.05.200).
8. Silver D., Huang A., Maddison C., Guez A., Sifre L., Driessche G., Schrittwieser J., Antonoglou I.,Panneershelvam V., Lanctot M., Dieleman S., Grewe D., Nham J., Kalchbrenner N., Sutskever I., Lillicrap T., Leach M., Kavukcuoglu K., Graepel T., Hassabis D. Mastering the game of Go with deep neural networks and tree search. Nature. 2007:484–503.
9. Omid E., Netanyahu N., Wolf L. DeepChess: End-to-end deep neural network for automatic learningin chess. Proc. of ICANN 2016: 25th Intern. Conf. on Artificial Neural Networks, Barcelona, Spain. Springer LNCS. 2016; (9887):88–96.
10. Makarov I.M., Lohin V.M. Intellektual’nye sistemy avtomaticheskogo upravleniya [Intelligent automatic control systems]. Moscow: Fizmatlit; 2001: 578. (In Russ.)
11. Belov M.P., Chan D.H. Intelligent controller based on non-linear optimal control of robotic manipulators [Intellektual’nyj kontroller na osnove nelinejnogo optimal’nogo upravlenija robotami-manipuljatorami]. Izvestiya SPbGETU LETI. 2018; (9):76–86. (In Russ.)
12. Alvarado R., Valdovinos L., Salgado-Jimenez T., Gomez-Espinosa A., Fonseca-Navarro F. Neuralnetwork-based self-tuning PID control for underwater vehicles. Sensors (Basel). 2016: 16(9):898–903.
13. Kumar R., Srivastava S., Gupta Artificial Neural Network based PID controller for online controlof dynamical systems. Proc. of Sensors IEEE 1st Intern. Conf. on Power Electronics, Intelligent Control and Energy Systems (ICPEICES), Delhi, 2016. Available at: https://ieeexplore.ieee.org/document/7853092
14. Zribi A., Chtourou M., Djemel M. A new PID neural network controller design for nonlinear processes. Available at: http://arxiv.org/abs/1512.07529
15. Voevoda A.A., Romannikov D.O. Synthesis of regulators for multichannel systems using neural networks. Scientific Bulletin of NSTU. 2019; 4(77):7–16. DOI:10.17212/1814-1196-2019-4-7-16. (In Russ.)
Bibliography link:
Voevoda A.A., Romannikov D.O. The synthesis method of regulators for multichannel systems using neural networks // Computational technologies. 2020. V. 25. ¹ 3. P. 111-118
|
