Master Thesis, 30HP: Finding the UAV in a Haystack of Birds -- Deep Models for Target Classification in Radar Tracking
Teknik, data och digitalt · Data, AI och analys · Maskininlärning · Datavetenskap · Mjukvaruutveckling
I korthet
This Master Thesis project at Saab in Göteborg, Sweden, focuses on developing deep models for target classification in radar tracking, specifically distinguishing UAVs from birds using Bayesian methods. The internship offers students hands-on experience in a real-world challenge within air-situational picture quality and safety.
Ansvarsområden
- Explore Bayesian methods for target classification.
- Exploit behavioral differences between birds and UAVs to determine target class.
- Investigate prospects of likelihood-based classification using learned system models.
- Potentially learn filters to facilitate likelihood evaluation within a modest computational budget.
- Consider a ladder of system models: traditional kinematic models, models in deep latent spaces, and hybrid models.
- Address how to combine prior structure and model expressiveness to maximize object classification capabilities.
Krav
- Student in the end of technical master's education in Engineering Physics, Engineering Mathematics, Electrical engineering, Automation and Mechatronics or similar.
- Interest for advanced mathematics and numerical methods.
- Practical experience of deep learning.
- Must pass a security vetting.
Önskade kvalifikationer
- Advanced courses in mathematics, in particular Bayesian statistics, is meriting.
- Citizen of Sweden may be required for security clearance (implied).
Förmåner
- Opportunity to collaborate with experienced engineers and specialists.
- Gain invaluable practical experience.
- Make a tangible contribution to company growth and development.
- Support and guidance to translate theoretical knowledge into practical solutions.
#master thesis#deep learning#radar tracking#target classification#UAV#Bayesian methods