Semidefinite Relaxation for Large Scale MIMO Detection

The semi-definite relaxation (SDR) is a high performance efficient approach to MIMO detection especially for low modulation orders. We focus on developing a computationally efficient approximation of the maximum likelihood detector (ML) algorithm based on semi-definite programming (SDP) for MQAM constellations. The detector is based on a convex relaxation of the ML problem. A comparative analysis including the performance-complexity trade-off of the SDR and the lattice reduction (LR) aided linear MIMO detectors considering high number of antennas is carried out aiming to demonstrate the effectiveness of the SDR-based conventional and large scale MIMO detector. SDR-MIMO detector can provide a close, and under high order antennas cases, a better performance than the LR-aided linear MIMO detectors.