AP26103493 – Machine Learning-Driven Intelligent Surfaces for Green IoT in Future Wireless Networks

Objective of the project – To develop ML-driven optimization techniques for RIS to enable energy-efficient and scalable IoT networks in next-generation wireless communications. The project addresses challenges such as massive connectivity, energy efficiency, and resource allocation, improving network performance and supporting sustainable wireless systems.

Relevance: The relevance of this project is driven by the rapid growth of IoT technologies and the need for energy-efficient and scalable wireless communication systems for future generations. Existing networks face significant challenges, including massive device connectivity, energy limitations, and complex signal propagation conditions. The integration of Reconfigurable Intelligent Surfaces (RIS) with machine learning provides a promising solution to optimize system performance in real time. This approach enables improved signal propagation, enhanced energy harvesting, and more efficient resource allocation in IoT networks. The project is also aligned with emerging 6G technologies, which are still at an early stage, creating a timely opportunity for innovation and leadership. Therefore, the project is highly relevant as it supports technological advancement, sustainable communication systems, and the development of smart infrastructure and Industry 4.0 in Kazakhstan.

Scientific supervisor: Ph.D., Associate Professor, Zhamangarin Dusmat Samatuly

Expected and achieved results: The project successfully developed a theoretical framework for RIS-assisted communication in IoT networks, including the analysis of RIS principles and their impact on wireless communication quality. Key RIS parameters such as phase control, reflection coefficients, and signal optimization methods were thoroughly studied. Existing channel models were reviewed, and the role of RIS in improving coverage and reliability of IoT networks was clearly identified. The theoretical work also highlighted the main limitations of RIS technology and proposed potential solutions, forming a solid foundation for further modeling. A technical report for the first stage (WP1) was prepared, summarizing objectives, methodologies, intermediate results, and recommendations for further development. Preliminary evaluation of the RIS-based system model was conducted, including analysis of resource allocation mechanisms and system performance under different IoT scenarios. The model’s stability and functionality were verified, allowing identification of its strengths and weaknesses and the creation of an initial dataset for future simulations. An analytical model was also assessed, with validation of assumptions, mathematical relationships, and system stability under varying parameters. The analytical results were further verified through MATLAB simulations, confirming the accuracy of most model predictions and identifying areas requiring refinement. The comparison between analytical and simulation results demonstrated the model’s applicability to real-world scenarios. The project results were systematized and used to prepare a scientific manuscript, including structured data, visual materials, and key findings. Activities on dissemination and project management were successfully carried out, ensuring proper reporting, coordination, and compliance with timelines. Additionally, a project website was developed and launched to support information sharing and public access to results. Overall, the project achieved significant progress in both scientific development and organizational implementation.