Abstract
In the downlink of a multiuser multiple-input
multiple-output (MIMO) communication system, simultaneous
transmission to several users requires joint optimization of
the transmitted signals. Allowing all users to have multiple
antennas adds an additional degree of complexity to the
problem. In this paper, we examine the case where a single
base station transmits to multiple users using linear
processing (beamforming) at each of the antenna arrays. We
propose generalizations of several previous iterative
algorithms for multiuser transmit beamforming that allow
multiple antennas and multiple data streams for each user,
and that take into account imperfect channel estimates at
the transmitter. We then present a new hybrid algorithm
that is based on coordinated transmit-receive
beamforming, and combines the strengths of nonorthogonal
iterative solutions with zero-forcing solutions. The
problem of distributing power among the subchannels is
solved by using standard bit-loading algorithms combined
with the subchannel gains resulting from the zero-forcing
solution. The result is a significant performance
improvement over equal power distribution. At the same time,
the number of iterations required to compute the final
solution is reduced.