Blind Detection Based on Asymmetric Constellations for MIMO Systems

In this paper, we propose a blind detection method based on asymmetric constellations for multiple-input multiple-output (MIMO) systems. In block fading channels, the transmission information sequence is randomly and uniformly generated from the constellation set, which results in signals classification at the receiver. We exploit the system constrained Gaussian mixture model (SCGMM) to detect information sequence. Specifically, we exploit the a posteriori probabilities (APPs) calculated by the expectation-maximization (EM) algorithm to recognize signals. For blind detection, we develop an unsupervised learning scheme based on asymmetric constellations, which solves the permutation ambiguity problem by asymmetrically designing traditional constellations. To obtain robust initial centroids, we propose an asymmetric constellations-based initialization algorithm, which can reduce the estimated error and accelerate convergence of the SCGMM clustering. The simulation results show that the bit-error-rate (BER) performance of the proposed method can approach the performance with perfect channel state information (CSI) in the high signal-to-noise (SNR) regions.