Master's Thesis presentation, Albin Ekelund Karlsson and Samuel Sandelius

Password for online link: 704048

Abstract: A modern radar antenna can direct its energy electronically without inertia or the need for mechanically steering. This opens up several degrees of freedom such as transmission direction and illumination time, and thus also the potential to optimise operation in real-time. A long range surveillance radars solve the trade-off between searching for new targets and tracking known targets. This optimisation is often rule-based. In this thesis we show that reinforcement learning can outperform such rule-based approaches for a simulated radar.