Paper Title
Simplified Computation Of Log-Likelihood Ratio Values For High-Order Modulations

Abstract
This paper proposes a method to approximately compute log-likelihood ratio (LLR) values with reduced complexity. Conventionally, the maximum-likelihood (ML) method is used as the optimum LLR computation method in terms of reception performance. However, the ML method has a drawback that its computational complexity increases exponentially with the modulation order of the system. Although this complexity issue may not significantly affect the reception time with low modulation orders, when high order modulations such as 4096-quadrature amplitude modulation (QAM) are used the ML method may not be used due to its enormous amount of computations. In this respect, this paper proposes an approximation to compute the LLR values. The proposed method requires only M one-dimensional (1D) Euclidian distances for one modulated symbol while the ML method needs 2M 2D distances which may not be implementable with high order modulation. As a cost of the complexity reduction, the proposed method shows an acceptable performance degradation. Keywords- High Order Modulation, Log-likelihood ratio, Maximum-Likelihood Estimation.