Generalized Shifting-Based PAPR Reduction Architecture for Discrete Multitone Systems

In this paper, a generalized shifting-based peak-to-average power ratio (PAPR) reduction architecture is presented for discrete multitone (DMT) systems. The limitation of the state-of-the-art scheme is that it resorts only to the local extrema from subsets in calculating the compensation offset, being inherently prone to accuracy degradation. To overcome the drawback, the proposed generalized offset enables new design options that utilize multiple global extrema, mitigating the PAPR further. In addition to the generalization, a low-complexity bilateral sorter, which is the key component of the entire architecture to identify multiple extrema, is developed by simplifying and hybridizing mergesorters. Since the generalized offset not only offers the new options but also encompasses all the previous expressions of the offsets in the literature, it successfully unifies all dc shifting-based PAPR reduction schemes, and exempts us from comparing them and choosing one.