Abstract :
[en] Low earth orbit (LEO) satellite communication systems have attracted extensive attention due to
their smaller pathloss, shorter round-trip delay and lower launch cost compared with geostationary
counterparts. In this paper, the downlink transmit design for massive multiple-input multiple-output
(MIMO) LEO satellite communications is investigated. First, we establish the massive MIMO LEO
satellite channel model where the satellite and user terminals (UTs) are both equipped with the uniform
planar arrays. Then, the rank of transmit covariance matrix of each UT is shown to be no larger than
one to maximize ergodic sum rate, which reveals the optimality of single-stream precoding for each
UT. The minorization-maximization algorithm is used to compute the precoding vectors. To reduce the
computation complexity, an upper bound of ergodic sum rate is resorted to produce a simplified transmit
design, where the rank of optimal transmit covariance matrix of each UT is also shown to not exceed
one. To tackle the simplified precoder design, we derive the structure of precoding vectors, and formulate
a Lagrange multiplier optimization (LMO) problem building on the structure. Then, a low-complexity
algorithm is devised to solve the LMO, which takes much less computation effort. Simulation results
verify the performance of proposed approaches.
Scopus citations®
without self-citations
30