Peak-to-Average Power Ratio Optimization For Multicarrier Communication Systems

Multicarrier communication technology has received a great deal of attention for its potential to achieve high data rate transmission over multipath fading channels and efficiently utilize the available spectrum. So far, orthogonal frequency division multiplexing, representing the multicarrier communication technology, has become a key technology of the fourth generation mobile communications and been successfully utilized in many wireless communications systems. MIMO OFDM, the combination of the multiple input multiple output and OFDM technology can increase the system capacity and improve spectrum efficiency effectively without increasing the bandwidth and the transmit power. It has been adopted for many wireless communication standards such as IEEE 802.11 (WLAN) and IEEE 802.16 (WiMAX). In MIMO OFDM system, beamforming is performed independently on each subcarrier to mitigate the effect of fading in wireless communication channels and to achieve maximized signal-to-noise ratio and the system performance can be further improved.However, the high peak-to-average power ratio (PAPR) of the transmitted signal is one of the major constraints for future development of the OFDM technology, especially in beamforming MIMO OFDM system, where the PAPR problem may become more severe due to the beamforming precoding. High PAPR necessitates the use of high power amplifier with large dynamic range to avoid out-of-band radiation and signal distortions, which would eventually result in low power efficiency of the power amplifier and high investment cost. Especially, for base station equipped with massive amounts of antennas, it is a paramount importance to reduce the PAPR problem.In this thesis, we firstly analyze the peak-to-average power problem in multicarrier systems, then based on the status of this problem, we concentrate our efforts on the research of the peak power optimization in OFDM, beamforming MIMO OFDM and massive antennas beamforming OFDM systems. The main contributions are summarized as follows:(1) Clipping-based tone reservation method is an efficient method to reduce the PAPR of the OFDM systems. The objective is to find the optimal peak-canceling signal to approximate the clipping noise and when the peak-canceling signal is added to the transmitted signal, the peak power would be reduced. Firstly, to amplify the amplitude of the peak-canceling signal in traditional clipping-based tone reservation, a novel scaling factor optimization function is proposed. Secondly, to improve the PAPR reduction performance, which is limited by the fixed clipping level in traditional clipping-based tone reservation, a novel cost function to jointly optimize the clipping level and the peak-canceling signal is proposed. The obtained optimal clipping level can not only render the clipping noise to be well approximated by the reserved tones, but also be small enough to deeply suppress the peaks. In addition, a new cyclic method to optimize the peak-canceling signal and the clipping level is presented with the insight that when optimizing the reserved tone values, the time-domain signal to be thresholded itself depends on the peak-canceling signal instead of being constant, as assumed by other methods. Compared to the the state-of-the-art clipping-based TR algorithms, the proposed system significantly improves the PAPR reduction performance with similar computation complexity.(2) The weighted tone reservation method is another efficient way to reduce the PAPR of the OFDM systems, where a weighted least squares optimization is performed to obtain the peak-canceling signals and the weights are based on the amplitude of the transmitted signal. This will ensure that the peaks of the clipping noise get better approximated and result in high suppression of the peak portion. However, the weighted tone reservation has two main drawbacks:the first is the arbitrarily selecting of the weights results in limited optimization solution; the second is the peak power of the transmitted signal is reduced without considering the probability of the potential power regrowth, thus the PAPR performance of weighted least squares tone reservation is limited. We firstly thoroughly study the optimal weight selection and provide the general guidelines. Secondly, we propose a new weighted tone reservation cost function by adding a well-designed penalty term that seeks to deeply suppress the peak regrowth of the transmitted signal in the low energy portion while remain the approximation of the peak portion. In addition, based on the derivation of the optimal clipping level and the peak-canceling signal, an iterative weighted tone reservation is proposed to further improve the PAPR reduction performance.(3) In beamforming MIMO OFDM system, adjusting the weights of the beamformer is a recently proposed and efficient way to reduce the PAPR. So far, the existing beamformer adjustment method is short of the intensive study of the received signal-to-noise ratio (SNR) and the relationship between the PAPR reduction and the SNR maximization, therefore it reduces the PAPR by heuristically amplifying the amplitude of the beamformer at the subcarriers with large SNR, and due to the heavy SNR degradation at some subcarriers, the method would increase the bit error rate of the OFDM systems. In this thesis, we formulate this problem of both lowering PAPR and maximizing the receive SNR in a new framework through singular value decomposition. This framework better elucidates the relationship between PAPR reduction and SNR maximization than previous efforts. Based on it, a SNR maximization PAPR reduction beamforming scheme is proposed and PAPR can be lowered significantly with SNR closing to the optimal one.(4) At present, massive antennas technology is considered as a key technology in 5G mobile communication system. With massive antennas system, we think of a base station with a few hundred antennas simultaneously serves a set of users to increase energy efficiency and spectrum usage. To avoid signal distortion, costly linear power amplifier will be the problem for the massive antennas system in practice use. Since the computational complexity of the traditional PAPR reduction techniques increases with the number of antennas at the base station, we focus on in this thesis low-complexity PAPR reduction methods in massive beamforming Multi-antenna OFDM system. The insight is on the observation that the arbitrary phase rotation applied to the beamformer can be used to reduce the PAPR while maintain the optimal SNR. Based on it, we propose a phase rotation low-complexity beamforming scheme. Experimental results show that the proposed beamforming scheme offers deeply PAPR suppression with low computational complexity and while remains the optimal SNR.