This paper proposes an efficient bit-loading algorithm for Multiple-Input- Multiple-Output transmission over spatially correlated channels with perfect Channel State Information at both link ends. Based onwater-filling principle, the algorithm does not assume the well-known gap approximation and accounts for the spatial correlation effects through the initial bit allocation. The result is the optimal distribution of bits and power over the eigenmodes for multiple antenna systems. We report the performance evaluations in the presence of spatial correlation in order to illustrate the efficiency and the robustness of our bit-loading algorithm in realistic environments. The simulation results demonstrate that our algorithm leads to significant improvements. In addition, these results show that our algorithm achieves the maximal diversity order, as well as exhibiting a high extra coding gain.
