Low Complexity And High Performance Sphere Detection Technique For MIMO Communication Systems

In literature many authors reported that in multiple input multiple output (MIMO) detection methods sphere-decoding (SD) is one of the efficient detection technique. In general, Euclidean norm is used for traversing a tree in sphere detection. By using the proposed method circuit complexity can be reduced with negligible performance loss. This is due to the reduction in the number of nodes visited in the critical path of the search tree. This paper presents the performance of complexity of 4x4 MIMO with 16-QAM modulation. The performance is compared with maximum likelihood detection (MLD) and L
²
norm with the help of MATLAB simulation results. It was observed that the complexity is reduced when mean of real multiplications and mean of real additions are calculated. Proposed L∞ sphere detection shows better performance as the number of transmitting antennas increased. The tree pruning behavior of proposed SD is aggressive than the previously reported ones. Proposed L∞ sphere detector follows new approach of radius selection for reducing the computational complexity as compared to the L
²
sphere de-coding algorithm and with the ideal ML decoder. It was also observed that using the infinity-norm reduces the hardware complexity of sphere detector considerably at only a minor performance loss.