Implementation of interaction between collaborative robots using federated learning in the form of digital twins

نوع: Type: Thesis

مقطع: Segment: masters

عنوان: Title: Implementation of interaction between collaborative robots using federated learning in the form of digital twins

ارائه دهنده: Provider: Sepehr Rezaei

اساتید راهنما: Supervisors: Dr. Hatam Abdoli

اساتید مشاور: Advisory Professors: Dr. Mahdi Abbasi

اساتید ممتحن یا داور: Examining professors or referees: Dr. Hassan Khotanlou - Dr. Reza Mohammadi

زمان و تاریخ ارائه: Time and date of presentation: 2025

مکان ارائه: Place of presentation: 30

چکیده: Abstract: The expansion of the Internet of Things (IoT) and the connection of billions of smart devices have introduced several challenges, including high energy consumption, limited computational resources, data security, and communication stability. Federated Learning (FL), as a distributed approach in machine learning, addresses these challenges by enabling data processing at the source without the need to transfer data to a central server. This approach reduces issues such as privacy violations and heavy network traffic, while offering a suitable foundation for IoT development by preserving data privacy and minimizing communication overhead. This study proposes an innovative method for collaborative robot interaction in IoT environments. The method utilizes collaborative learning within the framework of a digital twin to optimize energy consumption, reduce communication overhead, and enhance security in the learning process. The use of the MQTT communication protocol, due to its high efficiency in low-bandwidth environments and features such as data encryption, authentication, and low energy consumption, has played a significant role in improving network performance and communication security. Furthermore, the proposed model improves network stability and reduces energy consumption through the implementation of a sleep/wake mechanism and clustering techniques, minimizing reliance on a central server. Experimental results demonstrate that the proposed method outperforms conventional approaches in reducing energy consumption, improving learning time, enhancing communication security, and optimizing network performance. Among these achievements, energy optimization stands out as the most significant accomplishment, with the proposed model achieving a 34.6% reduction in energy consumption. This achievement highlights the effectiveness of the proposed algorithm in reducing energy costs and enhancing network sustainability.

فایل: ّFile: Download فایل