Half‑time seminar on 5G with Maxim Friesen
Maxim Friesen, PhD-student in a collaboration between TW-OWL, Germany and Mid Sweden University.
Welcome to a half-time seminar with Maxim Friesen, PhD-student in a collaboration between TW-OWL, Germany and Mid Sweden University.
Supervisors: Prof. Mikael Gidlund, Dr. Sarder Fakhrul Abedin and Prof. Juergen Jasperneite.
External opponent of half-time seminar: Dr Filip Barac, 6G Radio Standardization Expert at Ericsson AB
Seminar Title:
Towards Sustainable 5G Backhauling: A Deep Reinforcement Learning Framework for Energy-Efficient Mobile IAB Operation
Seminar Abstract:
The deployment of 5G and beyond (5G+) networks relies heavily on densification, where Integrated Access and Backhaul (IAB) offers a flexible wireless solution to traditional backhauling challenges. However, energy consumption in the Radio Access Network remains a significant operational cost, particularly in emerging mobile and battery-operated IAB systems. These systems are designed for dynamic environments such as industrial automation, mining, or emergency response, where fixed power infrastructure is unavailable or impractical. To address the challenge of energy efficiency in these mobile IAB networks, this research focuses on minimizing power consumption through the dynamic management of node power states. A Deep Reinforcement Learning (DRL) framework is proposed to learn optimal node activation policies that adapt to changing network conditions, including traffic loads and link quality. The approach is integrated using a comprehensive system-level simulator, leveraging the Sionna library for modeling the physical layer wireless channels and a rule-based low-level scheduler to manage backhaul transmissions. The platform is designed for training DRL agents and evaluating their ability to balance energy savings with network Quality of Service in realistic mobile IAB scenarios.