On Peak-To-Average Power Ratios In Multicarrier Communications

Multicarrier modulation (MCM) has been included in many communication standards and is promising for the future high speed wireless communications for its robustness against inter-symbol interference (ISI) caused by the frequency-selective property of nonideal channels and achievement of channel capacity with simple implementations in nonideal channels. However, an inherent drawback of MCM, high peak-to-average power ratio (PAPR) for a large subcarrier number, limits the use of MCM in communication systems. The aim of this dissertation is to analyze the statistical properties of PAPR in Multicarrier communications and effects of PAPR on system performance, to improve some existing techniques for PAPR reduction, and to probe some new, practical and effective PAPR reduction methods. The main contributions of this dissertation are as follows.A two-dimensional random walking model and a one-dimensional random walking model are proposed. Although they are very complicated, the two models can be used to precisely analyze the statistical properties of MC signals with any subcarrier number, modulation, and power allocation. A fast computational method using FFT is also proposed to simplify the analysis based on the random walking model.Based on the principle of selected mapping (SLM) method two novel approaches, successive interleaving SLM (SI-SLM) and short span interleaving SLM (SSI-SLM), are proposed for PAPR reduction. The SI-SLM uses an interleaver to produce different frames representing the same information by successive interleaving operations and transmits the most favorable one as regards to PAPR, which is simple in implementations compared to conventional interleaving based schemes. The SSI-SLM uses a convolution-like interleaver to produce different frames for SLM, which costs negligible memory and results in very short time delay. Both of the two novel approaches are very effective to depress the PAPR of high level modulated signals.Some new partial transmit sequence (PTS) based methods are proposed. These methods including binary tree based search, multi-initial iteration, trellis search and coded sequence try method are trade-offs between the PAPR reduction performance and the complexity of optimization for PTS combination. The...