Near-Capacity MIMO Subspace Detection

This paper presents a low complexity detection algorithm for Multiple-Input Multiple-Output (MIMO) systems. The novel subspace approach involves QR Decomposition (QRD) or Cholesky decomposition to triangularize the effective channel matrix so that several (possibly overlapping) groups of data streams can be detected separately. Link layer simulations show that the ergodic MIMO capacity can be closely approached by exploiting an Overlapped Subspace Detection (OSD) strategy. The OSD algorithm offers a scalable performance/complexity trade-off between Zero-Forcing (ZF) and Maximum a posteriori Probability (APP) detection. The proposed algorithm can be straightforwardly applied to large MIMO systems, as prevalent with recent advances in the field such as network MIMO and Coordinated Multi-Point (CoMP) transmission and reception.