Low-Complexity Non-Data-Aided SNR Estimation for Multilevel Constellations

Empirical-distribution-function (EDF)-based non-data-aided signal-to-noise ratio (SNR) estimation methods are effective for multilevel constellations but require many matching operations between reference cumulative distribution functions and EDF. To reduce resource consumption and improve real-time performance, we propose a low-complexity method by simplifying the matching test method. The proposed method modifies the matching operation from two-dimensions (2-D) to 1-D, reducing the number of matching operations and memory resource. Compared to the moment-based methods and the existing EDF-based reduced-complexity methods, the proposed method has the lowest complexity and provides better estimation performance in the medium-to-high SNR range.