CTH05-6: Hybrid-Selection/Equal-Gain Combining for Non-Constant Modulus Signals

In this paper, we study the performance of predetection hybrid-selection/equal-gain combining (HS/EGC), in which a subset of branches with the largest signal-to-noise ratios (SNRs) are selected and combined per EGC rule. To enable HS/EGC diversity for non-constant modulus (NCM) modulation formats, we propose the relevant HS/EGC receiver structure and decision variable. By deriving the moment generating function (MGF) of the HS/EGC output signals and with the help of the Parseval´s theorem, we develop a general analytical framework to evaluate the error and outage probabilities of pre-detection HS/EGC for a large class of modulation formats and versatile system and channel parameters. Simulation results verify the validity of our proposed receiver structure and the accuracy of our analysis. Performance comparison with HS/MRC shows that HS/EGC achieves the same diversity order as HS/MRC. Some other insightful findings including the effect of the subset combining loss are also provided.