Distributed Downlink Precoding and Equalization in Satellite Swarms

In this paper, we propose a novel approach for downlink transmission from a satellite swarm towards a ground station (GS). These swarms have the benefit of much higher spatial separation in the transmit antennas than traditional big satellites with antenna arrays, promising a massive increase in spectral efficiency. The resulting precoder and equalizer have very low demands on computational complexity, inter-satellite coordination and channel estimation. This is achieved by taking knowledge about the geometry between satellites and GS into account. For precoding, each satellite only requires its angle of departure (AoD) towards the GS and it turns out that almost optimal data rates can be achieved if the satellites transmit independent data streams. For the equalizer, the GS requires only knowledge about the angles of arrival (AoAs) from all satellites. Based on the underlying geometrical channel approximation, the optimal inter-satellite distance is obtained analytically. We show, that, by choosing a proper inter-satellite distance, the proposed low-complexity approach achieves the theoretical upper bound in terms of data rate. Furthermore, a novel approach to increase the robustness of the proposed precoder and equalizer against imperfect AoD and AoA knowledge is proposed, by exploiting the statistics of the estimation error.