Beam Direction Optimization for Next-Generation GEO Satellite Networks.

This paper develops a beam direction optimization framework for next-generation GEO satellite networks. The objective is to meet traffic demands at user locations. Given beam-pointing directions, the downlink of the GEO satellite is a vector broadcast channel that consists of a single transmitter and multiple distributed ground users. We characterize the downlink channel matrix for the multibeam satellite network by using an array factor formula for uniform planar arrays. We obtain a necessary and sufficient condition dependent on the downlink channel matrix to provision traffic demands by meeting given SINR targets at user locations. Utilizing the necessary and sufficient conditions, we formulate a joint beam direction and power optimization problem to attain target SINRs which uses minimum total power. Our results demonstrate that analog beamforming with optimized beam shifts can achieve an SINR gain of 8 dB when compared to analog beamforming without beam direction optimization. It also offers a spatial multiplexing advantage of 90 km by enabling simultaneous provisioning of user locations in close proximity within the same frequency band. When compared to hybrid beamforming, our scheme can achieve an SINR gain of 2 dB.