Enhanced control of a two-wheeled gyroscopic combat robot with variable CMG disk speeds using LQR-Based neural network

Gökhan Çetin; Yasin Oğuz; Faruk Ünker

doi:10.17714/gumusfenbil.1785098

TR EN

LQR tabanlı sinir ağı kullanılarak değişken CMG disk hızlarına sahip iki tekerlekli jiroskopik savaş robotunun gelişmiş kontrolü

Öz

Bu çalışma, derin öğrenme ile Doğrusal Kuadratik Regülatör (LQR) kontrolünü bütünleştirerek iki tekerlekli jiroskopik bir savaş robotunun kararlılığını ve uyarlanabilirliğini artırmayı amaçlayan hibrit bir kontrol yaklaşımı önermektedir. Doğaları gereği kararsız yapıda olan iki tekerlekli robotlar, özellikle sistem dinamiklerini önemli ölçüde etkileyen değişken Kontrol Momenti Jiroskobu (CMG) disk hızları altında dengeyi korumakta zorluk yaşayabilmektedir. LQR, bir maliyet fonksiyonunu en aza indirerek etkin bir dengeleme sağlasa da disk hızlarındaki değişimlere bağlı olarak performansı düşebilmektedir. Bu sorunu aşmak için, 800 ila 5000 aralığında aralığındaki disk hızları için optimize edilmiş 42001 önceden hesaplanmış LQR katsayıları üzerinde bir Yapay Sinir Ağı (YSA) eğitilmiştir. Eğitilen YSA, gerçek zamanlı olarak uygun katsayıları tahmin ederek dinamik koşullar altında uyarlanabilir kontrol sağlamaktadır. Simülasyon sonuçları, hibrit YSA-LQR denetleyicisinin, klasik LQR’nin 4–6 saniye içinde dengesini kaybettiği durumlarda bile 3000 N’ye kadar geri tepme kuvvetleri altında kararlılığı koruduğunu göstermektedir. Savaş robotu için tasarlanmış olmasına rağmen, önerilen çerçeve; arama-kurtarma, gözetleme ve karmaşık ortamlarda otonom gezinme gibi daha geniş robotik uygulamalar için ölçeklenebilir bir çözüm sunarak robotikte akıllı kontrolün gelişimine katkıda bulunmaktadır.

Anahtar Kelimeler

Jiroskopik kontrol , Hibrit kontrol sistemi , Doğrusal ikinci dereceden regülatör (LQR) , Sinir ağı , Kararlılık ve uyarlanabilirlik , İki tekerlekli robot

Enhanced control of a two-wheeled gyroscopic combat robot with variable CMG disk speeds using LQR-Based neural network

Abstract

This study proposes a hybrid control approach to improve the stability and adaptability of a two-wheeled gyroscopic combat robot by integrating deep learning with Linear Quadratic Regulator (LQR) control. Due to their inherently unstable nature, two-wheeled robots face challenges maintaining balance, especially when Control Moment Gyroscope (CMG) disk speeds vary, which significantly affect system dynamics. While LQR provides effective stabilization by minimizing a cost function, its performance degrades when disk speeds change. To overcome this limitation, a Neural Network (NN) was trained on 42001 precomputed LQR gain sets optimized for CMG disk speeds ranging from 800 to 5000 rad/s. The trained NN predicts appropriate coefficients in real time, enabling adaptive control under dynamic operating conditions. Simulation results demonstrate that the hybrid NN-LQR controller maintains stability under recoil forces up to 3000 N, where the conventional LQR loses balance about 4–6 seconds. Furthermore, the proposed controller reduces energy consumption by 21.6% during steady-state operation within the first 5 seconds and 10.6% under recoil conditions within the first 3 seconds compared to classical LQR. Although developed for a combat robot, the proposed framework offers a scalable solution for broader robotic applications such as search and rescue, surveillance, and autonomous navigation in complex environments, contributing to the advancement of intelligent control in robotics.

Keywords

Gyroscopic control , Hybrid control system , Linear quadratic regulator (LQR) , Neural network , Stability and adaptability , Two-wheeled robot

Kaynakça

Atac, E., Yildiz, K., & Ülkü, E. E. (2021). Use of PID Control during Education in Reinforcement Learning on Two-Wheel Balance Robot, Gazi University Journal of Science Part C: Design and Technology, 9 (4), 597-607. https://doi.org/10.29109/gujsc.955562
Åström, K. J. & Murray, R. M. (2010). Feedback Systems: An Introduction for Scientists and Engineers, Princeton University Press, New Jersey.
Çetin, G., & Ünker, F. (2024a). Gyroscopic precession control for maneuvering two-wheeled robot recoil stabilization. Journal of Field Robotics, 41(4), 991–1005. https://doi.org/10.1002/rob.22305
Çetin, G., & Ünker, F. (2024b). Control moment gyroscope recoil stabiliser including LQR controller for two-wheeled robot. International Journal of Heavy Vehicle Systems, 31(6), 782–797. https://doi.org/10.1504/IJHVS.2024.142258
Figurski, J., & Rybak, P. (2007). Fighting means quality and military vehicle load. Journal of KONES Powertrain and Transport, 14(2), 125–132.
Haarnoja, T., Zhou, A., Abbeel, P., & Levine, S. (2018). Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor, In 6th International Conference on Learning Representations (ICLR 2018), (pp. 1-14). https://doi.org/10.48550/arXiv.1801.01290
Gautam, P. (2016). Optimal control of inverted pendulum system using ADALINE artificial neural network with LQR. In 2016 International Conference on Recent Advances and Innovations in Engineering (ICRAIE) (pp. 1–6). https://doi.org/10.1109/ICRAIE.2016.7939523
Kim, S. & Kwon, S. (2017). Nonlinear Optimal Control Design for Underactuated Two-Wheeled Inverted Pendulum Mobile Platform, IEEE/ASME Transactions on Mechatronics, 22 (6), 2803-2808. https://doi.org/10.1109/TMECH.2017.2767085
Lillicrap, T. P., Hunt, J.J., Pritzel, A., Heess, N., Erez, T., Tassa, Y., & Silver, D. (2016). Continuous control with deep reinforcement learning, In 4th International Conference on Learning Representations (ICLR 2016), (pp. 1-14). https://doi.org/10.48550/arXiv.1509.02971
Nakamura-Zimmerer, T., Gong, Q., & Kang, W. (2021). QRnet: Optimal regulator design with LQR-augmented neural networks. IEEE Control Systems Letters, 5(4), 1303–1308. https://doi.org/10.1109/LCSYS.2020.3034415

Nghi, H. V., Nhien, D. P., & Ba, D. X. (2022). A LQR neural network control approach for fast stabilizing rotary inverted pendulums. International Journal of Precision Engineering and Manufacturing, 23, 45–56. https://doi.org/10.1007/s12541-021-00606-x
Uddin, N., Nugroho, T. A., & Pramudito, W. A. (2017). Stabilizing Two-wheeled robot using linear quadratic regulator and states estimation, In 2nd International conferences on Information Technology, Information Systems and Electrical Engineering (ICITISEE), (pp. 229-234). https://doi.org/10.1109/ICITISEE.2017.8285501
Ünker, F. (2023a). Recoil stabilization of two-wheeled robot through control moment gyroscope (CMG). In 2023 5th International Congress on Human-Computer Interaction, Optimization and Robotic Applications (HORA) (pp. 1–6). https://doi.org/10.1109/HORA58378.2023.10156663
Ünker, F. (2022a). Proportional control moment gyroscope for two-wheeled self-balancing robot. Journal of Vibration and Control, 28(17–18), 2310–2318. https://doi.org/10.1177/10775463211009988
Ünker, F. (2022b). Oscillation control of two-wheeled robot using a gyrostabilizer. Gazi University Journal of Science Part C: Design and Technology, 10(3), 547–557. https://doi.org/10.29109/gujsc.1062497
Ünker, F. (2023b). Gyroscopic suspension for a heavy vehicle. International Journal of Heavy Vehicle Systems, 30(2), 201–210. https://doi.org/10.1504/IJHVS.2023.132315
Ünker, F. (2022c). Angular momentum control for preventing rollover of a heavy vehicle. International Journal of Automotive Science and Technology, 6(2), 135–140. https://doi.org/10.30939/ijastech..1085111
Ünker, F. (2020). Tuned gyro pendulum stabilizer for control of vibrations in structures. International Journal of Acoustics and Vibration, 25(3), 355–362. https://doi.org/10.20855/IJAV.2020.25.31632
Yan, A., Chen, Z., Dong, C., & He, K. (2021). Attitude balance control of two-wheeled robot based on fuzzy reinforcement learning, Systems Engineering and Electronics, 43 (4), 1036–1043. https://doi.org/10.12305/j.issn.1001-506X.2021.04.21

Ayrıntılar

Birincil Dil

İngilizce

Konular

Akıllı Robotik , Kontrol Mühendisliği , Kontrol Mühendisliği, Mekatronik ve Robotik (Diğer)

Bölüm

Araştırma Makalesi

Yazarlar

Gökhan Çetin ^*
0000-0002-7960-1217
Türkiye

Yasin Oğuz
0000-0002-0324-0515
Türkiye

Faruk Ünker
0000-0002-9709-321X
Türkiye

Yayımlanma Tarihi

6 Şubat 2026

Gönderilme Tarihi

16 Eylül 2025

Kabul Tarihi

30 Ocak 2026

Yayımlandığı Sayı

Yıl 2026 Cilt: 16 Sayı: 1

DOI

https://doi.org/10.17714/gumusfenbil.1785098

IZ

https://izlik.org/JA76HS56FX

APA

Çetin, G., Oğuz, Y., & Ünker, F. (2026). Enhanced control of a two-wheeled gyroscopic combat robot with variable CMG disk speeds using LQR-Based neural network. Gümüşhane Üniversitesi Fen Bilimleri Dergisi, 16(1), 110-125. https://doi.org/10.17714/gumusfenbil.1785098

e-ISSN: 2146-538X Yayın Modeli: Sürekli Yayın Atıf Dizinleri: TR Dizin , SCOPUS

LQR tabanlı sinir ağı kullanılarak değişken CMG disk hızlarına sahip iki tekerlekli jiroskopik savaş robotunun gelişmiş kontrolü

Öz

Anahtar Kelimeler

Enhanced control of a two-wheeled gyroscopic combat robot with variable CMG disk speeds using LQR-Based neural network

Abstract

Keywords

Kaynakça

Ayrıntılar

Birincil Dil

Konular

Bölüm

Yazarlar

Yayımlanma Tarihi

Gönderilme Tarihi

Kabul Tarihi

Yayımlandığı Sayı

DOI

IZ

Kaynak Göster