Analysis of PAPR Reduction in OFDM Systems

Abstract

This thesis considers the long-standing problem of Peak-to-Average-Power Ratio (PAPR) reduction in OFDM systems. Due to the nonlinear amplifiers commonly used, OFDM signals often will be distorted when they are transmitted. Many methods have been proposed before to reduce the PAPR: block coding, clipping, and insertion are the most common. They all change the signal by adjusting the signal or its input data. In this thesis, 3 methods that take a different approach are discussed: Gradient Descent, Newton’s Method, and Single Peak Suppression. These methods aim to move the constellation points in order to lower the peak power. Different OFDM system parameters regarding the number of subcarriers and the constellation size are considered and the results show that all three methods can reduce the PAPR. Single Peak Suppression is the fastest and the simplest method overall. Gradient Descent can produce well suppressed signal at a medium complexity. While Newton’s Method can reach low PAPR in the fewest iterations, its computational complexity for each iteration makes it the least efficient overall to carry out.