Optimum rotation parameters for single and multiuser signal space diversity employing spatial diversity

Abstract

Signal space diversity (SSD) is an efficient scheme to obtain diversity without expanding bandwidth. In the literature, the best rotation matrices of SSD are obtained by the minimum product-distance metric. In this work, a new metric is employed, which relates directly the rotation matrix to the mean bit error rate (BER) for systems with or without an antenna array at the receiver. Single user per channel, quadrature amplitude modulation (QAM) and flat fading channel scenarios are considered. Besides that, a maximum likelihood detector (MLD) is employed at the receiver, which complexity is evaluated. Orthogonal matrices that achieve the best rotation angles are obtained via exhaustive computational search, using the new metric for a D-dimensional space and ϒ antennas at the receiver. Simulation results show that the best rotation matrices obtained with the new metric are superior in performance in relation to the matrices obtained with the minimum product distance metric. It is also verified that the best rotation matrices change with the number of dimensions, modulation order and number of antennas. In addition, a multiuser scenario is evaluated by considering different rotation parameters for each user. The best rotation matrix of the single user scenario could be used for all users in the multiuser scenario, without significant performance reduction.