Low-Complexity PTS Schemes Using Dominant Time-Domain Samples in OFDM Systems

Abstract

In orthogonal frequency division multiplexing (OFDM) systems, high peak-to-average power ratio (PAPR) of OFDM signals is one of the most important problems. The high PAPR of OFDM signals causes serious nonlinear distortions in process of passing through high power amplifier (HPA). These distortions have a effect on in-band distortion and out-of-band radiation, which result in bit error rate degradation of received OFDM signals and interference in adjacent channel, respectively. In order to solve the PAPR problem of OFDM signals, various PAPR reduction schemes have been proposed. This dissertation includes research results on a kind of the PAPR reduction schemes, called the partial transmit sequence (PTS) for the OFDM systems. As a solution to the PAPR problem in OFDM systems, the PTS scheme is a fairly suitable scheme due to its PAPR reduction performance and distortionless characteristics. The PTS scheme generates several candidate OFDM signals to represent an original OFDM signal and selects one with the lowest PAPR among them for transmission. However, a serious problem in the PTS scheme is high computational complexity, which is mainly required to generate and process the candidate OFDM signals. In this dissertation, in an effort to reduce its computational complexity, new PTS schemes are proposed using dominant time-domain samples of OFDM signals. Dominant time-domain samples is a small number of samples of OFDM signals used to estimate PAPRs of candidate OFDM signals efficiently. In the first part of this dissertation, low-complexity PTS schemes are proposed using new selection methods of dominant time-domain samples. The proposed selection methods of dominant time-domain samples are based on selection methods of candidate samples in candidate OFDM signals. These methods select dominant time-domain samples with reduced computational complexity. The dominant time-domain samples selected by the proposed methods are used to estimate PAPRs of candidate OFDM signals with high accuracy. Therefore, the proposed low-complexity PTS schemes can achieve the optimal PAPR reduction performance with considerably reduced computational complexity. In the second part of this dissertation, improved PTS schemes are proposed to lower the computational complexity of previous PTS schemes further while maintaining high performance of PAPR reduction. Similar with the PTS schemes proposed in the previous part of this dissertation, the improved PTS schemes utilize dominant time-domain samples and candidate samples. However, they use more efficient methods, which select the candidate samples by adaptive method or multi-stage method to select dominant time-domain samples. Therefore, the improved PTS schemes reduce computational complexity further while maintaining the optimal PAPR reduction performance. The proposed PTS schemes in this dissertation use efficient methods to select dominant time-domain samples and thus they reduce the computational complexity considerably compared to previous PTS schemes. In addition, they achieve the optimal PAPR reduction performance, which is equivalent to that of the conventional PTS scheme with the low complexity. Due to the high performance and low complexity, they are fully expected to be used in the practical implementation of OFDM systems.