On the Performance of Cognitive Satellite-Terrestrial Networks

We investigate the performance of a multi-beam cognitive satellite terrestrial network in which a secondary network (mobile terrestrial system) shares resources with a primary satellite network given that the interference temperature constraint is satisfied. The terrestrial base stations (BSs) and satellite users are modeled as independent homogeneous Poisson point processes. Utilizing tools from stochastic geometry, we study and compare the outage performance of three secondary transmission schemes: first is the power constraint (PCI) scheme where the transmit power at the terrestrial BS is limited by the interference temperature constraint. In the second scheme, the terrestrial BSs employ directional beamforming to focus the signal intended for the terrestrial user, and in the third, BSs that do not satisfy the interference temperature constraint are thinned out (BTPI). Analytical approximations of all three schemes are derived and validated through numerical simulations. It is shown that for the least interference to the satellite user, BTPI is the best scheme. However, when thinning is not feasible, PCI scheme is the viable alternative. In addition, the gains of directional beamforming are optimal when the terrestrial system employs massive multiple-input-multiple-output transceivers or by the use of millimeter wave links between terrestrial BSs and users.