Abstract: The method for decoding a received signal in a multiple input/multiple output system uses QR-decomposition of the linear channel matrix, but then applies a non-Euclidean norm during tree traversal. Two separate hardware units, namely an MCU and a MEU, art provided for concurrent operation. The MCU determines a next child node, while the MEU determines next best parent nodes on the previously processed tree levels, which makes it possible to retrace the path to a next starting node without investing dedicated processing steps (e.g., cycles). On each tree level, the possible coordinates are grouped into several circular sets in the complex plane, and a series of decision boundaries is calculated for each set that allows a quick evaluation of the optimum coordinate in each set.

Abstract: The method for decoding a received signal in a multiple input/multiple output system uses QR-decomposition of the linear channel matrix, but then applies a non-Euclidean norm during tree traversal. Two separate hardware units, namely an MCU and a MEU, art provided for concurrent operation. The MCU determines a next child node, while the MEU determines next best parent nodes on the previously processed tree levels, which makes it possible to retrace the path to a next starting node without investing dedicated processing steps (e.g., cycles). On each tree level, the possible coordinates are grouped into several circular sets in the complex plane, and a series of decision boundaries is calculated for each set that allows a quick evaluation of the optimum coordinate in each set.