Two‐dimensional sorting algorithm for high‐throughput K‐best MIMO detection

Abstract

This brief presents a novel two-dimensional parallel sorting algorithm for high-throughput K-best detectors utilised in multiple-input multiple-output (MIMO) systems. The proposed sorting algorithm enhances the throughput by sorting a data set in parallel and avoids the relatively long latency of the traditional algorithms. This is especially important in MIMO systems utilising high-order modulation schemes. The authors used the sorting algorithm in a K-best detector with variable K values at different layers of the search tree, which improves the bit error rate performance and reduces the computational complexity significantly. The detector using the sorting algorithm is designed and implemented in TSMC 90-nm CMOS technology for 4 × 4 64-QAM MIMO systems. Operating at 200 MHz, the detector's throughput is 1.2 Gbps. Its equivalent gate count is 182K.