Background
For suporting such requirements, two main deployment modes have been standardized by 3GPPP, i.e., Non-Standalone (NSA) and Standalone (SA). The prime difference between the two modes is that 5G NSA is supported by an existing 4G core which acts as a control plane, while 5G SA is completely independent, thus, employing its own 5G core for all data and control operations. In the current deployment phase, the majority of the operators adopt NSA, since it is less complex and costly to implement. However, the co-dependece between the two technologies introduce several challenges mainly in terms of performance, coverage, and network management (e.g., gNB placement strategies, handovers, network configurations, etc.) that require further analysis.
To explore the above aspects, we carry out a large-scale measurement campaign on 5G NSA commercial cellular networks for four major operators in Rome, Italy, during a period of approximately 2 months (i.e., beginning April - late May). During that period, we performed a plethora of passive and active measurements using a commercial measurement setup provided by Rohde and Schwarz (R&S). The collected dataset will allow the students to conduct several analyses on network coverage, RAN deployment, and end-user performance leveraging artificial intelligence (AI) and machine learning (ML) techniques. To support such tasks, they will get access to our open-source large-scale dataset along a detailed documentation that they can leverage for exploring several aspects of the network from multiple angles.
Expected results and learning outcomes
- Combine your technical and creative skills;
- Gain experience in applying AI and ML in a large-scale dataset of mobile measurements;
- Enrich your profile with networking knowledge;
- Provide valuable insights on the coverage, deployment and performance of the rapidly growing 5G networking ecosystem;
- Possibility to publish the results and attend a top international conference.
Conditions
We expect that you:
- have been admitted to a master's program in MatNat@UiO - primarily PROSA
- take this as a long thesis
- participate actively in the weekly SINLab meetings
- are present in the lab and collaborate with other students and staff
- are interested in and have some knowledge in R, Python, or Matlab programming
- have basic to good knowledge around the different concepts of mobile networks
- are willing to share your results on Github
- include the course IN5060 in the study plan, unless you have already completed a course on classical (non-ML) data analysis
Contacts
- Konstantinos Kousias, UiO, email: konstako@ifi.uio.no
- Ozgu Alay, UiO, email: ozgua@ifi.uio.no
- Carsten Griwodz, email: griff@ifi.uio.no