About the opportunity
Ericsson is a world-leading provider of telecommunications equipment and services, supporting mobile and fixed networks in more than 180 countries. Our global and innovative environment gives students the opportunity to work on challenging research problems with real-world impact.
The State of Charge (SOC) is a key parameter in battery operation. Accurate SOC estimation is essential for managing the operating conditions of Radio Base Stations (RBSs) and enabling them to support Smart Grid services.
However, estimating the SOC accurately throughout a battery’s lifetime remains challenging. In this thesis, you will investigate energy-efficient AI-based methods for predicting battery SOC and optimising battery operation in RBS power systems. The project aims to improve battery performance and energy utilisation while identifying local operating constraints and performance limitations.
The thesis is intended for one student, and the scope can be adapted to the candidate’s interests and background.
What you will do
During the thesis, you will:
- Conduct a literature review of relevant concepts and algorithms for analysing battery data from RBS power systems.
- Investigate and test suitable machine-learning and AI algorithms for SOC estimation.
- Select, model, and implement appropriate techniques for different scenarios within the chosen use case.
- Develop an end-to-end prototype or proof of concept.
- Evaluate the resulting model using relevant battery and operational data.
- Analyse the model’s accuracy, computational requirements, energy efficiency, and potential limitations.
- Identify opportunities for improving battery management and energy utilisation in RBSs.
The skills you bring
We are looking for a highly motivated student who enjoys challenging research work and is eager to propose and develop innovative ideas. You have:
- MSc studies in Computer Science, Mathematics, Physics, Engineering, or a related field.
- Excellent programming skills in Python.
- Good knowledge of machine learning, including deep learning and unsupervised learning.
- An interest in data analysis, prediction, and battery systems.
- Experience with machine-learning libraries and frameworks such as TensorFlow, Keras, PyTorch, Scikit-Learn, or Spark.
- Strong analytical and problem-solving skills.
- An interest in developing end-to-end prototypes and practical concepts.
- The ability to work independently while collaborating effectively with others.
- Fluency in English.
Knowledge of battery systems, time-series analysis, Docker containers, orchestration systems, or telecommunications is considered an advantage.
When applying, please include a transcript of records showing your completed courses and grades.
