Spatial diversity techniques provide the receiver with multiple replicas of the transmitted signal and help combat fading effects over wireless links. This talk provides an overview of recent results on precoding schemes for multi-user wireless downlink channels when multiple antennas are employed at both the transmitter and receiver. The proposed MIMO precoding schemes differ in several respects from previous work. First, each user can have multiple receive antennas and the transmitter uses multiple beams instead of a single beam in order to transmit multiple symbols simultaneously to each user. This generalization is useful for space-time coding schemes. We select the beamforming weights from the channel space of all users excluding the desired user, while at the same time maximizing the link capacity. Second, several power allocation schemes are derived for the cases of a homogeneous network (where all users have the same transmit power), a non-homogeneous network (where users can have different transmit powers), as well as for networks with a uniform quality of service (i.e., same throughput for all users). Third, we design precoders that minimize the interference power caused by a user on all other users, as opposed to forcefully nulling the interference.
The resulting schemes relax the traditional constraint on the number of transmit and receive antennas. Specifically, the performance criterion is related to maximizing a so-called signal-to-leakage-plus-noise ratio (SLNR) for all users simultaneously. The results suggest dynamic antenna scheduling strategies for downlink MIMO communications, where a subset of the receive antennas at certain users are selectively disabled. The largest value that the SLNR can achieve is shown to be equal to the maximum eigenvalue of a certain random matrix combination, and the probability distribution of this eigenvalue is characterized in terms of a Whittaker function. The result shows that increasing the number of antennas at some users can degrade the SLNR performance at other users. This fact suggests an antenna scheduling scheme that leads to significant improvement in terms of SNR outage probabilities.