Join our Team

About this opportunity:
Master Thesis: Building Loss and Indoor Traffic Classification

This is an opportunity for a Master of Science student to work with network data for an estimation problem using statistical and/or machine-learning approach together with data-driven modeling of building loss using ML techniques for a state-of-the-art network simulator.


5G is expected to become the dominant mobile access technology by subscription in 2028. About 300 communications service providers (CSPs) globally now offer 5G services, of which about 50 have launched 5G Standalone (5G SA). At the same time, most of the data is consumed or generated indoors as we tend to spend more time indoors. For 4G/LTE, about 70-90% of the data for the outdoor base-station is consumed or generated indoors depending upon the deployment scenarios (dense-urban, sub-urban etc.). 5G spectrum ranges from low band to midband to mm Wave. Higher frequencies of 5G may have higher building penetration loss but may be good for higher capacity. Standalone 5G or NR-SA are usually deployed with a low-band FDD (e.g., 700 MHz) for coverage and mid-band TDD (3.5 GHz) for capacity with massive-MIMO systems.


What you will do:
The objective of the master thesis is to estimate what ratio of the traffic at the outdoor 5G base-stations is consumed or generated indoors in different frequency bands with focus specially on mid and high TDD bands. This can be done using the network data through statistical and/or machine learning techniques. The obtained results will help us in understanding indoor performance and deployment of various frequency bands specially mid and high TDD bands in a standalone 5G network in a better way.


At the same time, as indoor coverage is influenced by the building penetration loss, proper modeling of building loss in network simulator is critical to understand indoor performance. Another objective of the thesis would be to implement data-driven methodology, e.g., using various attributes of buildings from publicly available open street map, to develop ML model to estimate building loss.


The skills you bring:

• Excellent grades
• Fluent in English, both written and spoken
• Good MATLAB/Python skills
• Good communications skills
• You are a self-motivated and positive person.
• Experience with Statistical and Machine Learning approach is a bonus.

Extent: 1 student
Work location: Stockholm, Kista
Preferred Starting Date: Q1 2025, please enclose a transcript in your application and state when you can start.