A Transmit Preprocessing Technique for Downlink Multi-User MIMO System Using Subspace Decomposition Approach

Based on the research on architecture of multi-user MIMO channel, we introduce the concept of accompanying subspace of the channel, and conclude several linear methods into the problem that how to produce preceding matrix on this subspace. And we provide an iterative algorithm to get optimal solution on this subspace based on the criterion of maximum channel capacity. By computer simulation, we analyze the performance of our algorithm and compare it with other methods. Results show that our method has good performance: (1) Compared with channel inversion and block diagonalization, it greatly increases outage capacity, especially when Pout is small; (2) Ergodic capacity grows linearly with the increase of the number of transmitting antennas, and is much bigger than that of other methods; (3) It can better deal with near-far effect and keep spatial division multiplexing gain.