Downlink spectrum sharing of heterogeneous communication systems in LEO satellite networks

We study downlink spectrum sharing in low earth orbit (LEO) satellite networks, where wide beam coverage and side lobe beam make it hard to manage mutual interference between satellites. As a first step, we consider a two-satellite system of control satellite and interference satellite, which share the same frequency spectrum range and are located close enough to have interference each other. We investigate resource allocation strategy for the control satellite to maximize its throughput performance while satisfying interference constraints. Assuming that the control satellite has no prior knowledge about the behavior of the interference satellite, we develop online frequency allocation algorithms that successfully manage the interference by employing a learning-based design with UCB index. Under the proposed algorithms, we can maximize throughput of the control satellite while constraining interference at the interference satellite as well as at the control satellite, without any direct information exchange with the interference satellite. Through simulations, we demonstrate that the proposed schemes achieve high throughput performance satisfying interference constraints.