Background of the thesis work
At Swedspot we work with a rich variety of smart vehicles all connected to the internet over a cellular network. A smart connected vehicle can have hundreds of various sensors that produce data in abundance. Due to response time, bandwidth capacity, and storage needs it makes little sense to transfer all of this data to a cloud backend system for further computing.
An alternative approach would be to put more intelligence in the edge equipment. By using sensor fusion to aggregate sensors we could define a meaningful high-level events interface. These events could be used by both local and remote algorithms.
Area of thesis research
- Conduct a literature study on approaches to create sensor fusion and local event detection systems
- Define possible architectures, APIs, and best practices on how to build a flexible sensor fusion system
- Make an implementation of the system on Swedspots Ezentric platform that demonstrates the usefulness of the system
- Write a report, suggest recommendations for sensor configurations, and identify topics for future research
Suitable background of students:
This thesis is suited for students with a Master of Science in Computer Science or a similar education. You will possess good communication skills in English, shows a high level of initiative, and is self-driven.
Application deadline: 10:th of December
Suggested start: Jan 2023
Number of students: Preferably the thesis is done by at least two persons