Improving the Reliability of the K-Best algorithm for MIMO detection with ordering

Abstract

It is well known that the order of the channel matrix columns has significant impact on a MIMO detector's performance in terms of the computational complexity, memory requirement, and/or the detection error rate. In our previous work, novel ordering schemes have been proposed to reduce the computational complexities and/or memory requirements of various maximum likelihood (ML) MIMO detectors. In this paper, we incorporate our ordering schemes with the K-Best detector, which is a near-ML detector and is particularly suitable for hardware implementation. Our simulation results show that our ordering schemes greatly improve the reliability of the K-Best detector. Given a fixed detection error rate, our ordering schemes either result in SNR gains or enable the usage of even smaller K, thereby leading to small area and power consumption and higher throughput for their hardware implementations.