Reduction of Peak-to-average Power Ratio in Orthogonal Frequency Division Multiplexing Radio Systems

Abstract

Orthogonal frequency division multiplexing (OFDM) has continued to be the most preferred signal-multiplexing scheme for the current and future standards for high-speed transmission of data over the wireless channel. This is due to its ability to provide a high spectral efficiency and to combat both intersymbol interference and multipath fading. However, an OFDM signal tends to have a high peak-to-average power ratio (PAPR) when the number of subcarriers is large. Subsequently, the high power amplifier in the transmitter nonlinearly amplifies the signal resulting in-band and outof- band radiations in the output of the device. In turn, these radiations give rise to bit-error rate (BER) degradation and adjacent channel interferences. Therefore, it is necessary to reduce the PAPR to avoid these two effects. This thesis proposes five new PAPR reduction methods for OFDM systems with a contribution to the class of additive signal methods. Three of the proposed methods follow the conventional tone-reservation (TR) approach; these are second-order cone program TR (SOCP-TR), linear program TR (LP-TR) and iterative re-weighted least squares TR (IRLS-TR). The remaining two methods, namely low-complexity signal addition (LCSA) and low-complexity additive signal mixing (LCASM), extend the transmit time to cater for the transmission of nonzero samples of the peak-reduction signal. The SOCP-TR and LP-TR are optimal schemes, while the remaining three are suboptimal and are preferable for practical realization due to their fast convergence rates and low computational complexities. The suboptimal methods can achieve similar or better PAPR reductions than the optimal schemes. From the perspective of the reservation of PAPR reduction resources, in either frequency or time domain, the IRLS-TR and LCSA methods offer the best performances. For example, with 5% of the transmit resources reserved in a system with 256 subcarriers, the IRLS-TR and LCSA methods achieve reductions of 5.84 and 5.69 dB, at the expense of only a small change in transmit power of 0.36 and -0.10 dB, respectively. Additionally, the five proposed methods reduce PAPR without degrading the BER. The proposed methods also perform better than other methods proposed in literature. In addition to offering new PAPR reduction methods, this thesis has established that the type of subcarrier modulation and signal structure, real or complex, insignificantly affect the amount of PAPR reduction.