Reduced complexity for the radius adaptive K-best algorithm

Abstract

The detection of multiple-input multiple-output (MIMO) system is an important issue. The radius adaptive K-Best (RAKB) algorithm is proposed for the detection of the MIMO system, it decomposes the searching tree into several subbranches and provides similar bit-error-rate (BER) performance to the K-Best algorithm. But the complexity of the RAKB algorithm is still very high, it will pay an additional arithmetic at low signal-to-noise ratio (SNR). In this paper, we improve the RAKB algorithm to simply the complexity. The improved algorithm only searches the preinstall sub-branch at low SNR. Simulation results show that the improved algorithm achieves similar performance to the K-Best algorithm with a significantly reduced complexity, in terms of the number of visited nodes and computational time.