Constellation Collaborated Non-linear Orthogonal Space-Time Block Codes with Fast Maximum Likelihood Detection

Abstract

In this paper, constellation collaborated nonlinear orthogonal space–time block codes (OSTBCs) for wireless communication systems are considered. Using the recently developed unique factorization of signals, several uniquely factorable constellation pairs (UFCPs) generated from cross quadrature amplitude modulation (QAM) constellations are chosen for transmission through all the transmitter antennas, with one random symbol acting as a coordinator to control all the symbols embedded in each coding matrix. The coding gains of collaborated orthogonal codes are calculated, and some optimal constellation pairs with maximum coding gains are presented, including orthogonal codes with different antennas and UFCPs with different groups. Despite the fact that the proposed code is nonlinear, a fast maximum likelihood (ML) decoding algorithm is developed, and its complexity is analyzed in detail. By comparing the decoding complexity of conventional Alamouti code, the ML algorithm, and the fast ML decoding algorithm of collaborative Alamouti code, it can be observed that the complexity of fast ML algorithm is still linear in terms of the square root of the cardinality of the cross QAM constellations with no performance loss. Computer simulations are performed to verify the performance of collaborative OSTBCs and the fast ML Algorithm under different conditions, and the results demonstrate that the average error rate of the collaborative nonlinear orthogonal codes outperforms that of the conventional OSTBCs.