APPLICATION OF METHODS FOR INCREASING THE EFFICIENCY OF NEURAL NETWORK ANALYSIS OF HETEROGENEOUS SYSTEMS-ON-CHIP IN ROBOTICS TASKS

M. L. Abdulvadudov1, K. L. Vakhidova2, S. M. Nogamirzaev2

About the authors

M. L. Abdulvadudov1, K. L. Vakhidova2, S. M. Nogamirzaev2

Ногамирзаев Сулиман Мовлаевич – старший преподаватель кафедры «Технологии машиностроения и транспортных процессов» Грозненский государственный нефтяной технический университет имени академика М.Д. Миллионщикова, адрес: 364051, иеченская Республика, г. Грозный, пр-т им. Х.А. Исаева, 100, Область научных интересов: компьютерное моделирование систем и процессов

Download PDF

UDC: 004.032.26
DOI: 10.24412/2413-7383-2025-4-39-163-172
Language: Russian

Abstract: The aim of this work is to study methods for improving the efficiency of neural network algorithms on heterogeneous RoboDeus systems-on-a-chip (SoCs), aimed at application in robotic systems. An analysis of the RoboDeus SoC architecture, including the Elcore50 DSP core, is presented, and optimization of convolutional neural networks is demonstrated using the YOLO V5s object detector as an example. The proposed approaches to parallelizing computations, efficient use of memory, and accumulator registers enable significant performance gains when executing computer vision tasks in real time. Experimental results show that the us e of multi -core processing based on RoboDeus provides a frame rate of up to 33.3 FPS for the task of detecting objects in 640×640 images. Prospects for integrating optimized neural network algorithms into autonomous robotic systems for navigation, environment recognition, and interaction with objects are discussed.

Keywords: heterogeneous computing systems, RoboDeus, YOLO V5s, convolutional neural networks, DSP core, instruction set architecture (ISA), system on a chip (SoC).

References

1. JSC Scientific and Production Center "ELVEES." Electronic resource: https://elvees.ru (date accessed: 24.10.2025).
2. Galushkin, A. I. Neural Network Theory: a textbook for universities / edited by A. I. Galushkin. Moscow: IPZhR, 2000. Book 1. 416 p.
3. Osovsky, S. Neural Networks for Information Proce ssing. // Translated from Polish by I.D. Rudinsky. Moscow: Finance and Statistics, 2002. 344 p.
4. Rostovtsev, V. S. Artificial Neural Networks [Electronic resource]: a textbook for universities. 2nd ed. reprinted. St. Petersburg: Lan, 2021. 216 p. – Access mode: https://e.lanbook.com/book/160142 (date accessed: 20.08.2022).
5. Tsurikov A. N. Modeling and training of artificial neural networks [Electronic resource]: textbook. manual. Rostov-on-Don: RGUPS, 112 p. Access mode: https://e.lanbook.com/book/1406 10 (date accessed: 20.08.2022).
6. A Heterogeneous Multi -Core System -on-Chip for Energy Efficient Brain Inspired Computing / Antonio Pullini, Francesco Conti, David Rossi, Michael Gautschi / ResearchGate, 2017. – 6 p.
7. Towards AI. Official website. Electronic resource: https://towardsai.net (date accessed: 24.10.2022).
8. Kunzman, D.M. (2011). Programming Heterogeneous Systems. International Symposium on Parallel and Distributed Processing Workshops. doi:10.1109/IPDPS.2011.377
9. Fedorov, M. V. (Corr esponding Member of the Russian Academy of Sciences) New Methods of Machine Learning: Using Fundamental Knowledge to Solve Practical Problems: [online abstract report] [Electronic resource] / M. V. Fedorov // Program of the II All -Russian School -Seminar on Artificial Intelligence and Big Data in Technical, Industrial, Natural, and Social Systems. – National Center for Physics and Mathematics (NCPM). – Date: November 25 -29, 2024. – (date of access: 04.04.2025).
10. Chudina, O.V. Theory and Practice of Heat T reatment of Metals: a teaching aid for a multimedia publication / O. V. Chudina, G. V. Gladova, A. V. Ostroukh. Moscow: MADI, 2013. 64 p.
11. Sergeev, A. P. Introduction to Neural Network Modeling: a tutorial / A. P. Sergeev, D. A. Tarasov; edited by A. P. Sergeev. Yekaterinburg: Ural University Publishing House, 2017. 128 p. ISBN 978 -5-7996-2124-7
12. Khlivnenko, L. V. Practice of Neural Network Modeling: a tutorial for universities / L. V. Khlivnenko, F. A. Pyatakovich. 2nd ed. reprinted. Saint Petersbur g: Lan, 2021. 200 p.: ill. Text: direct. ISBN 978.5.8114.8264.1
13. Vichugov V.N. Modified Gradient Algorithm for Training Radial Basis Neural Networks. Bulletin of Tomsk Poly technic University. Management, Computer Science, and Information Technology. 20 09. Vol. 315. No. 5, pp. 150 -152
14. Sidorov D.I. Novikova O.I. Application of Machine Learning Methods for Predictive Equipment Diagnostics. Problems of Artificial Intelligence, 2023, No. 1(28), pp. 23 -34.
15. Kozlov S.V. Frolova M.P. Optimization of En ergy Consumption of Industrial Installations Using Neural Networks. Problems of Artificial Intelligence, 2021, No. 4 (27), pp. 67 -78.
16. Belov I. A. Orlova T. K. Adaptive automatic control systems using hybrid algorithms. Problems of Artificial Intelligence, 2022, No. 2 (29), pp. 18 -29.
17. Nikitina L. V. Pavlov V. G. Modeling and analysis of complex technical systems under uncertainty. Problems of Artificial Intelligence, 2023, No. 4 (31), pp. 78 -90.
18. Fedorov K. A. Semenova A. A. Digital twins as a basis for increasing the efficiency of production lines. Problems of Artificial Intelligence, 2021, No. 1 (24), pp. 55 -66.
19. Grigoriev P. S. Medvedeva E. V. Intelligent data analysis in process parameter monitoring tasks. Problems of Artificial Intellig ence, 2022, No. 4 (31), pp. 34 -45.
20. Titov V.M. Vetrova N.S. Development of control algorithms based on reinforcement learning for industrial robots. Problems of Artificial Intelligence, 2023, No. 2 (29), pp. 61 -72. Применение методов повышения эффективности нейросетевого анализа…
21. Orlova S.K. Krylov D.V. Methods for increasing the stability of automatic control systems under disturbances. Problems of Artificial Intelligence, 2021, No. 3 (26), pp. 41-52.
22. Nurulin Yu. R. Skvortsova I. V. Nurulin M. Yu. Naumov A. S. Development of human -machine interfaces for

Issue: 4(39)'2025
Section: Роботы, мехатроника. Робототехнические системы
How to cite: M. L. Abdulvadudov1, K. L. Vakhidova2, S. M. Nogamirzaev2. APPLICATION OF METHODS FOR INCREASING THE EFFICIENCY OF NEURAL NETWORK ANALYSIS OF HETEROGENEOUS SYSTEMS-ON-CHIP IN ROBOTICS TASKS // Problems of Artificial Intelligence. - 2025. - № 4 (39). - P. 163-172. - https://paijournal.guiaidn.ru/en/2025/4(39)-14.html

Back to issue 4(39)'2025