A Novel Orthogonal Experimental Design Based QR Subspace MIMO Detection

Orthogonal experimental design (OED) is a well-known method in statistical experiment which aims to reduce the testing complexity. This paper tries applying OED to reduce the computational complexity in Multiple-input Multiple-output (MIMO) detection. Several times of sorted QR decomposition (SQRD) algorithm are applied firstly to get the diverse solutions. Then a subspace is constructed with the diverse results. Inside the subspace, OED shows its potential and value in reducing the computational complexity compares with the exhaustive method. The proposed QR-OED MIMO scheme provides nearly 8 dB gain compared with the traditional sorted QRD while the minimal computational complexity needed is only 3.7% compared with exhaustive method for the 6×6 MIMO system and 4-QAM modulation scheme.