Enhanced MIMO LMMSE Turbo Equalization

We propose an enhanced linear minimum mean square error (LMMSE) turbo equalization scheme for multiple-input multiple-output (MIMO) communication systems. The turbo equalizer employs soft interference cancellation (SIC), where tentative soft decisions of the interfering symbols are subtracted from the received samples before the LMMSE filtering. The enhanced LMMSE turbo equalization with SIC introduces two improvements over existing methods. First, unlike existing methods that completely rely on the a priori soft decisions at the equalizer input for SIC, the new scheme performs SIC by continuously updating the soft decisions with the a posteriori information of the symbols that have just been equalized. The a posteriori soft decision has a better quality than the a priori soft decision, thus it leads to a better SIC performance. Second, since the a posteriori soft decision of an equalized symbol will affect the SIC operation of all the subsequent symbols to be equalized, the order in which the symbols are equalized and detected plays a critical role on the overall performance. We propose a reliability-based detection ordering scheme, where symbols with more "reliable" soft input will be equalized before those less "reliable" symbols. The reliability information is extracted from the symbol a priori probability, which is a unique byproduct of turbo equalization and can be obtained with minimum extra cost. Simulation results demonstrate that the proposed scheme achieves considerable performance gain over the conventional turbo equalization methods.