Relaxed-bound K-best sphere detection for differential unitary space-time modulation

Large scale multiple input multiple output (MIMO), commonly referred to as massive MIMO, is one of the most important physical layer disruptions in network evolution that would lead to significant increase in spectral and energy efficiency envisioned for 5G networks. However, the computational complexity of massive MIMO system becomes the key problem for its practical implementation. As known, Differential Unitary Space-Time Modulation (DUSTM) has been proposed as an efficient modulation choice for such MIMO transmissions where channel state information (CSI) is not required at the receiver. However, traditional DUSTM implementation relies mainly on Maximum Likelihood Detection (MLD) achieving optimal performance, which is not easily compatible with massive MIMO systems due to its exponentially increasing complexity with increasing number of transmitting and receiving antennas. Therefore, in this paper, in order to overcome the aforementioned disadvantages, we propose a new detection algorithm named Relaxed-Bound K-Best Sphere Detection (RBKSD). Compared to the traditional K-Best Sphere Detection (KSD) algorithm, the main idea of our algorithm is to relax the bound K of the conventional KSD to be K + Γ, thereby achieving quasi-optimal performance with significantly reduced computational complexity. Simulation results confirm that the proposed RBKSD algorithm shows superior performance and reduced complexity compared to other notable schemes such as MLD, traditional sphere detection algorithm and the KSD.