Quantized modulation diversity for 64-QAM

Abstract

Communication through fading channels with conventional signal constellation has a limitation that the nearby points may lead to wrong symbol detection even if any one dimension fades. This limitation can be overcome by phase rotation of symbols where each symbol is rotated by a rotation matrix. In the new constellation, each symbol attains unique real and imaginary components which imply higher diversity order, and thus, it may result in higher noise margin. Though performance gain can be achieved through the selection of optimum rotation matrix, the high decimal precision of the rotated symbols demands larger memory support and processing effort due to their representation by unquantized real numbers. Quantization of the rotated symbols is a solution for this problem, but it may degrade the bit-error performance. This paper investigates the tradeoff between the bit-error performance of 64-QAM and the compromise in the precision of rotation.