Article 12320

Title of the article



Prokopyev Igor Vitalevich, doctor of technical sciences, leading researcher, Federal research center «Computer science and control» of RAS
(Dorodnitsyn computer center of the Russian Academy of Sciences) (bld. 2, 44 Vavilova street, Moscоw, Russia),
Sofronova Elena Anatolievna, candidate of technical sciences, associate professor, senior researcher, Federal research center «Computer science and control» of RAS (Dorodnitsyn computer center of the Russian Academy of Sciences) (bld. 2, 44 Vavilova street, Moscоw, Russia),

Index UDK





This paper presents the results of nonlinear model identification of an unmanned vehicle in the Gazebo simulator based on a neural network autoregressive model. The dynamic characteristics of the control object can vary significantly that complicates the problem. Data on the movement along the spatial path of an unmanned vehicle for the training sample was obtained at the Robotics Center of the FRC CSC RAS. The model parameters were found by the particle swarm optimization method. Using the identified model, real-time control methods were experimentally compared. Control was carried out on the basis of the PID controller, model predictive control  method and identifier based on artificial neural networks. In a comparative experiment, for a more accurate assessment of the identification method, the model was subjected to random disturbances, and the tracking path of the control object was significantly complicated. 

Key words

identification, unmanned vehicle, path tracking. 

 Download PDF

1. L'yung L. Identifikatsiya sistem [System identification]. Moscow: Nauka, 1991, 432 p. [In Russian]
2. Pupkov K. A., Egupov N. D., Trofimov A. I. Statisticheskie metody analiza, sinteza i identifikatsii nelineynykh sistem avtomaticheskogo upravleniya [Statistical methods for analysis, synthesis, and identification of nonlinear automatic control systems]. Moscow: MGTU im. Baumana, 1998, 562 p. [In Russian]
3. Kruglov V. V., Borisov V. V. Iskusstvennye neyronnye seti. Teoriya i praktika [Artificial neural network. Theory and practice]. Moscow: Goryachaya liniya – Telekom, 2002, 382 p. [In Russian]
4. Rutkovskaya D., Pilin'skiy M., Rutkovskiy L. Neyronnye seti, geneticheskie algoritmy i nechetkie sistemy [Neural networks, genetic algorithms, and fuzzy systems]. Moscow: Goryachaya liniya – Telekom, 2004, 452 p. [In Russian]
5. Paden B., Cáp M., Yong S. Z., Yershov D., Frazzoli E. IEEE Transaction on intelligent vehicles. 2016. DOI 10.1109/TIV.2016.257806.KorpusID:1229096
6. Fairchild C., Harman T. L. ROS Robotics By Example. Packt Publishing, 2016.
7. Gazebo. Available at:
8. Dar'ina A. N., Diveev A. I., Prokop'ev I. V. Voprosy teorii bezopasnosti i ustoychivosti system [Questions of the theory of security and stability of systems]. 2019, iss. 21, pp. 66–77. [In Russian]
9. Kingma D. P., Ba J. A. 3rd International Conference for Learning Representations. arXiv:1412.6980v8 [cs.LG]. San Diego, 2015, 15 p. Available at:
10. Lindfield G. Introduction to Nature-Inspired Optimization. Academic Press, 2017, pp. 49–68. Available at:
11. Diveev A. I., Pupkov K. A., Sofronova E. A. Vestnik RUDN, Ser.: Inzhenernye issledovaniya [RUDN Bulletin, Ser.: Engineering research]. 2011, no. 2, pp. 113–125. [In Russian]
12. Mehta B. R. Industrial Process Automation Systems. Butterworth-Heinemann, 2015, pp. 547–557. Available at:
13. Huang B., Kadali R. Dynamic Modeling, Predictive Control and Performance Monitoring: A Data-driven Subspace Approach, Lecture Notes in Control and Information Sciences. Springer, 2008, 241 p.
14. Pham D., Liu X. Neural Networks for Identification, Prediction and Control. Springer, 1995, 238 p.
15. Marquardt D. SIAM J. Appl. Math. 1963, no. 11, pp. 164–168.
16. Levenberg K. A Quart. Appl. Math. 1944, no. 2, pp. 164–168.
17. Broyden C. G. Journal of the American Mathematical Community. 1967, vol. 21, pp. 368–381.


Дата создания: 24.11.2020 14:47
Дата обновления: 24.11.2020 16:20