Subchannel allocation in multiuser multiple-input-multiple-output systems

Assuming perfect channel state information at the transmitter of a Gaussian broadcast channel, strategies are investigated on how to assign subchannels in frequency and space domain to each receiver aiming at a maximization of the sum rate transmitted over the channel. For the general sum capacity maximizing solution, which has recently been found, a method is proposed that transforms each of the resulting vector channels into a set of scalar channels. This makes possible to achieve capacity by simply using scalar coding and detection techniques. The high complexity involved in the computation of this optimum solution motivates the introduction of a novel suboptimum zero-forcing allocation strategy that directly results in a set of virtually decoupled scalar channels. Simulation results show that this technique tightly approaches the performance of the optimum solution, i.e., complexity reduction comes at almost no cost in terms of sum capacity. As the optimum solution, the zero-forcing allocation strategy applies to any number of transmit antennas, receive antennas and users.