Fast power minimization with QoS constraints in multi-user MIMO downlinks

In the downlink of a multi-user MIMO (multiple input multiple output) communication system where each user has an arbitrary QoS requirement, intelligent algorithms are needed to choose transmit vectors. Here we present a new method of choosing transmit vectors that minimizes total transmitted power. The approach is based on previous iterative interference balancing algorithms, but it is initialized by applying a "block-diagonalization" algorithm that helps improve convergence speed. When the channel supports multiple data streams per user, power is distributed among the data streams by bit-loading using the channel gains derived from the block-diagonalization step. The result is a solution which is not guaranteed to converge to the global optimum, but will reach a solution that is either optimal or near-optimal with high probability and at minimal computational cost.