Low-Complexity MIMO Detection Using Post-Processing SINR Ordering and Partial K-Best Search

Linear detectors such as zero-forcing (ZF) and minimum mean square error (MMSE) require only a small fraction of computational complexity compared to maximum likelihood (ML) detector. However, linear detections suffer from severe performance degradation. In this paper, we propose a novel detection scheme which obtains the initial symbol detection by MMSE detector and then perform symbol ordering by signal-to-interference-and-noise ratio (SINR). The MMSE detected symbols with higher SINR are retained as part of final solution and cancelled from the original received signals. The remaining symbols with lower SINR are detected by K-best algorithm, which selects K best nodes in each layer of the partial tree search. The small value of K is sufficient to achieve good performances, and therefore the extra computational complexity is minimal. Simulation results show the performance superiority of the proposed method compared to the conventional MMSE detection. Moreover, at the similar symbol error rates, the total number of nodes visited in the proposed approach is much smaller than the conventional K-best detection scheme.