Block Gauss-Seidel Method Based Detection in Vehicle-to-Infrastructure Massive MIMO Uplink

Vehicular ad hoc networks (VANET) have gained increasing interests due to the development of the intelligent transport systems (ITS), aiming to improving road safety, traffic efficiency, and providing in-vehicle entertainment. Meanwhile, the fast developing 5G cellular networks have brought innovative techniques to support the demand of ITS such as high rate communications, low latency and high energy efficiency. Massive multiple-input multiple-output (MIMO), as one of the key technologies in future 5G, is to deploy hundreds of antennas at base station, serving up to tens of users simultaneously in shared time-frequency resources. This technique, is attractive for the wireless vehicle-to-infrastructure (V2I) access for multiple vehicles on the road. However, in massive MIMO, the computational complexity is costly even with linear detection methods. The iterative methods, such as Gauss-seidel based signal detection method, are preferred as the computational complexity is low, and near-optimal system performance can be achieved. In this paper, we propose block Gauss-Seidel method based signal detection in V2I massive MIMO uplink transmission. The proposed scheme utilizes the properties of block diagonal matrix, and the Gauss-Seidel method is applied to each block. By doing that, the processing at each block can be paralleled, hence the new structure is much efficient for hardware implementation. In addition, we demonstrate that the system performance is quite close to the original Gauss-Seidel method but at low complexity and fast processing time.