About

• Join our Team About this opportunity: A sleeping cell is an unwanted condition where the cell is hanging, resulting in degraded, or no service, without alerting the operator about the ongoing situation.​ The operators are under pressure for better quality of service and network reliability. Sleeping cells are not possible to predict when and where a sleeping cell will occur.​ It is very important to quickly detect sleeping cells when they occur in the field and to perform automated recovery actions (e.g. cell restart).​ What you will do: In order to increase the accuracy of existing AI Powered Advanced Cell Supervision feature, there is a need to better understand if a cell is sleeping when the traffic is totally zero.​ There is a need to implement a solution that can accurately differentiate between an actual sleeping cell and a normal scenario (no connected users due to suburban area or during night time). The following steps are envisioned as part of the thesis work: Investigate efficient ways to collect historical data, with the requirements of low processing capacity usage, and limited data storage space. Investigate different machine learning models to predict if a real sleeping cell has been detected. Analyze results of predicted detections and evaluate performance of different machine learning models. The thesis will be concluded with a result presentation for the Ericsson research team. The skills you bring: This project aims at students in electrical engineering, computer science, computer engineering or similar. 1 student, 30 hp Ericsson AB, Kista Preferred start: Jan 2026