Radio scheduling for eXtended Reality (XR) in 5G Networks (with Ericsson Research)
Extended Reality (XR), which includes Cloud Gaming (CG) and Augmented/Virtual/Mixed Reality (AR/VR/MR) applications, is a key use case for 5G and beyond-5G systems. In order to satisfy XR requirements in terms of throughput and latency, several challenges are yet to be addressed. In particular, XR applications require heavy and nearly-periodic data transmission in both downlink and uplink, thus calling for significant enhancements in the scheduling and resource allocation mechanisms currently used at the radio access level.
Learning outcome
This thesis will explore the possibility of proposing and validating enhanced mechanisms for scheduling and resource allocation in 5G and beyond-5G radio access networks, in order to support XR services and meet quality of service and experience requirements. The investigation will leverage existing mechanisms for scheduling and resource allocation in 5G New Radio, such as dynamic scheduling, semi-persistent scheduling, and configured grant, aiming at enhancing and tailoring such mechanisms for XR traffic. The use of machine learning and artificial intelligence schemes will also be considered, in order to further optimize the proposed schemes.
Learning outcome
Good understanding of mobile communication systems (4G, LTE, 5G NR), practical experience with network simulators, ML-oriented Analytics.
Qualifications
Knowledge of programming languages (Matlab, Python), preferred knowledge of network simulators (e.g., ns-3, OMNeT++), preferred knowledge of Machine and Deep Learning schemes (e.g., Reinforcement Learning)