Thesis: Prediction of measurements in time series using Machine Learning i Järfälla

Saab Surveillance focus resides in finding efficient solutions for survaliance and desision support as well as systems that detects threats and protects the platform. The product portfolio consists of airborne, landbased, and naval radar systems. Futhermore, systems for signal inteligence and self protection are developed. There are also command and control systems for maratime and airtrafic control applications.

The proposed thesis project aims at evaluating different tracking and prediction Machine Learning algorithms based on time series data. In this case the time series is generic radar signals. They can be characterized as fix, stagger, and/or jitter.

You will be located at SAAB Surveillance (Järfälla), which will provide supervision concerning the development and provide necessary software tools.  

Assignment

The main objective with the thesis work is to apply Machine Learning algorithms for tracking and prediction of different generic radar signals that have been generated in a simulation. Trade-offs between speed, precision, and robustness shall be presented for selected Machine Learning algorithms (RNN, LSTM). The algorithms needs to be simulated and evaluated with different generic radar signals.

We believe that you are a motivated student with a strong interest in scientific computing and AI (e.g. deep learning, genetic algorithm). Furthermore MatLab and Python skills are essential and experience in object oriented programming and numerical methods is valuable. Prior knowledge of radar and RF environment is beneficial but not required. In addition you are fluent in spoken and written Swedish.

This position includes access to classified information, therefore you are required to conduct a security screening in accordance with the Protective Security Act and be approved for security clearance by the authority.