Noktasal Kontağa Sahip Zıplayan Robotta Duruş Anında Dengenin Sağlanması

Yıl 2024, Cilt: 36 Sayı: 2, 593 - 607, 30.09.2024

Halit Hülako

https://doi.org/10.35234/fumbd.1415814

Öz

Bu çalışmada, yer ile noktasal bir temas yüzeyine sahip olan zıplayan düzlemsel dinamik bir robotun zıplama sonrasında devrilmeden dengesini koruması incelenmiştir. Robotun dengesini sağlamak için Reaksiyon Tekerleği kullanılmıştır. Zıplama sonrasında yere iniş anı simüle edilmiş ve robotun yere temas ettiği anda dengesini koruyarak istikrarlı bir duruş sergilemesi için Genetik algoritma kullanarak Pekiştirmeli Öğrenme (PÖ) yapısına benzer bir Yapay Zeka Ajanı oluşturulmuştur. Bu yöntemle zıplayan robot, gerçek zamanlı olarak dengelenme misyonu için eğitilmiş ve ardından robotun dengeleme işlemini başarıyla öğrendiği gözlemlenmiştir. Zıplayan robotun dinamik modellemesi Matlab Simscape MultiBody (MSM) kullanılarak gerçekleştirilmiştir. Nümerik Simülasyona ait grafiksel sonuçlar, dinamik robot modelinin, zıplamadan sonra devrilmeden dengesini 5 saniye boyunca başarıyla sürdürdüğünü ve istenen düzeltmeleri yapabilme yeteneğini Yapay Zeka' nın katkısıyla sergilediğini göstermektedir.

Anahtar Kelimeler

Zıplayan robot, E. pekiştirmeli öğrenme, reaksiyon tekeri.

Maintaining Stability During Rest for Jumping Robot with Point Contact

Öz

This study investigates a planar modelled dynamic jumping robot with a point ground contact surface. The aim is to maintain the robot's balance after jumping without tipping over. To achieve this, a Reaction Wheel is employed for stabilization. The landing moment after a jump is simulated, and an Artificial Intelligence algorithm similiar to Reinforcement Learning (RL), which uses Genetic algorithm, is used to maintain a stable posture by maintaining balance at the moment the robot touches the ground. With this method, the jumping robot was trained for the stabilisation mission in real-time and then it was observed that the robot successfully learnt the stabilisation process. The jumping robot's dynamic modelling was carried out using Matlab Simscape MultiBody (MSM). The graphical results of the numerical simulation show that the dynamic robot model successfully maintains its balance for 5 seconds without tipping over after jumping and exhibits the ability to make the desired corrections with the contribution of Artificial Intelligence.

Anahtar Kelimeler

Jumping robot, E. reinforcement learning, reaction wheel.

Kaynakça

• Rui C, Kolmanovsky IV, McClamroch NH. Nonlinear attitude and shape control of spacecraft with articulated appendages and reaction wheels. IEEE Trans Automat Contr 2000; 45(8): 1455-1469.
• Shengmin G, Hao C. A comparative design of satellite attitude control system with reaction wheel. In: Proceedings - First NASA/ESA Conference on Adaptive Hardware and Systems, AHS 2006; pp. 359-362.
• Yang Y. Spacecraft Attitude and Reaction Wheel Desaturation Combined Control Method. IEEE Trans Aerosp Electron Syst 2017; 53(1): 286-295.
• Brown TL, Schmiedeler JP. Reaction Wheel Actuation for Improving Planar Biped Walking Efficiency. IEEE Transactions on Robotics 2016; 32(5): 1290-1297.
• Lee CY, Yang S, Bokser B, Manchester Z. Enhanced Balance for Legged Robots Using Reaction Wheels. In: 2023 IEEE International Conference on Robotics and Automation (ICRA); 2023; pp. 9980-9987.
• Graduate BB, Assistant R, Manchester Z. Rex Hopper: Design And Control Of A Monopod Hopper With Reaction Wheels.
• Zhang P, Wu Z, Dong H, Tan M, Yu J. Reaction-Wheel-Based Roll Stabilization for a Robotic Fish Using Neural Network Sliding Mode Control. IEEE/ASME Transactions on Mechatronics 2020; 25(4): 1904-1911.
• Gajamohan M, Merz M, Thommen I, D’Andrea R. The Cubli: A cube that can jump up and balance. In: IEEE International Conference on Intelligent Robots and Systems; 2012; pp. 3722-3727.
• Muehlebach M, Mohanarajah G, D’Andrea R. Nonlinear analysis and control of a reaction wheel-based 3D inverted pendulum. In: Proceedings of the IEEE Conference on Decision and Control, Institute of Electrical and Electronics Engineers Inc.; 2013; pp. 1283-1288.
• Brown TL, Schmiedeler JP. Reaction Wheel Actuation for Improving Planar Biped Walking Efficiency. IEEE Transactions on Robotics 2016; 32(5): 1290-1297.
• Han SI, Lee JM. Balancing and velocity control of a unicycle robot based on the dynamic model. IEEE Transactions on Industrial Electronics 2015; 62(1): 405-413.
• Trentin J, da Silva S, Ribeiro J, Schaub H. Inverted Pendulum Nonlinear Controllers Using Two Reaction Wheels: Design and Implementation. IEEE Access 2020; PP: 1.
• Hülako H, Yakut O. Control of Three-Axis Manipulator Placed on Heavy-Duty Pentapod Robot. Simul Model Pract Theory 2021; 108: 102264.
• Zhang S, Zhang H, Fu Y. Leg Locomotion Adaption for Quadruped Robots with Ground Compliance Estimation. Appl Bionics Biomech 2020; 2020: 8854411.
• Belascuen G, Aguilar N. Design, Modeling and Control of a Reaction Wheel Balanced Inverted Pendulum. In: 2018 IEEE Biennial Congress of Argentina (ARGENCON); 2018; pp. 1-9.