Active Constellation Extension for Peak Power Reduction Based on Positive and Negative Iterations in OFDM Systems

Abstract

Traditional active constellation extension (ACE) techniques iterate under a further and further away from decision boundary constraint to find distortions for peak-to-average power ratio (PAPR) reduction, which may stop the solution on suboptimal points because it's not permitted to go back when running into a suboptimum direction. In this paper, we present a novel ACE technique by iterating in both positive and negative directions, referring to distortions found in the last iteration. During iterations, optimization variations are changed from normally used extra distortions on the last estimates to the primitive OFDM signal, which can eliminate correlations between magnitudes and phases of complex distortions and finally give an analytic solution based on orthogonal projection. By making iterations run in positive and negative directions, this algorithm can find distortions to reduce PAPR more, compared with existing methods. Simulation results show that significant improvement can be achieved either for pure ACE or TR assisted ACE method, especially under higher-order modulation schemes.