Computationally efficient modulation detector with near optimal performance

Abstract

Maximum likelihood (ML) based modulation detector provides the optimal performance in the sense that the detection error probability is minimized, if no prior probability of candidate modulations is available at the modulation detector. However, the evaluation of the likelihood function requires prohibitively high computational complexity. This contribution deals with an approximation of the ML detector, which utilizes the special arrangement of square-formed quadrature amplitude modulation (QAM) schemes. Simulation results show that this approximated ML detector is able to provide near-optimal performance with moderate computational complexity.