Master Thesis: Generating realistic radar video data using AI and GAN

Vacancies 23 October 2019

Closing date





Åsa Mårtensson [C] +46 (10) 2165107
Ulf Näsström +46 (31) 7949051
Apply for job!

Master Thesis: Generating realistic radar video data using AI and GAN

Master thesis work at SAAB Surveillance Radar Engineering/X Innovation Lab 2020


Generative Adversarial Networks (GAN) is a class of machine learning systems where two neural networks compete against each other. The goal is to train a network to generate data to fool the other network into believing the generated data is real rather than artificial.

Developing modern radar systems is a long and complex process that is made even longer by the fact that simulating the signal environment in which the radar will operator is very difficult. Traditional methods of generating test data is too predictable and a system typically needs many hours of real world testing to find bugs and evaluate performance.

Master thesis work

Research questions

1.       Can GAN be used to generate radar data from a textual scenario specification?

2.       Can GAN be used to generate data in real time?

3.       Is the data generated of such quality that it can be used to test real radar application software?

4.       Is the data generated of such quality that it can be used to train other AI applications?


The goal of this master thesis work is to build a GAN that can generate raw radar data from a specification file. The generated data will then be used to test the software of the radar system and potentially also to train other AI to classify objects.

The work

The work will be divided into two parts. The first is a study of GAN to gain a deeper understanding of the technology and what is currently used in industry and the academic world. Based on this research, the students will then propose an implementation of GAN and proceed to implement this to demo the performance.


The work is expected to deliver the following:

a)       An overview of current research with regards to GAN.

b)      An overview of available tools and libraries that can be used to work with GAN at SAAB

c)       A demo of a GAN that can produce raw radar data.

d)      Suggestions on how to proceed, either with further studies or industrialization of the work done

What you will be a part of

Saab är ett globalt försvars- och säkerhetsföretag verksamt inom flyg-, land- och
marinförsvar, civil säkerhet och kommersiell flygteknik. Vi är 14 700 medarbetare och har verksamhet på alla kontinenter. Tekniskt är vi ledande inom många områden och en femtedel av våra intäkter går till forskning och utveckling.

Saab är också ett möjligheternas företag. Ett företag där vi ser mångfald som en tillgång och där du som medarbetare får stort ansvar och goda utvecklingsmöjligheter. Men också ett företag som respekterar varje människas behov av ett liv utanför arbetet.

Apply for job!