Research On Reducing Peak-to-Average Power Ratio Of OFDM Signals Based On Semideifnite Programming

Orthogonal Frequency Division Multiplexing (OFDM) is an efficient modulationtechnique. It has been applied to many wireless communication standards due to itsmerits of the high spectral efficiency and robustness against multipath propagation.OFDM signals, however, is with high peak to average power ratio (PAPR), whichresults in low efficiency of power amplifier in OFDM transmitter. To treat this problem,this thesis proposes a new approach to reduce PAPR values of OFDM signals based onthe semidefinite programming and semidefinite relaxation techniques, which aregradually mature in mathematical fields. And the simulation results are provided todemonstrate its effectiveness. The main contents in this thesis are:1. The effects of repeat clipping filtering, tone reservation and the second-ordercone programming for reducing PAPR values of OFDM signals are discussed.Simulation results of these methods are given to show their performances.2. The optimization model of minimizing EVM with PAPR constraints isproposed. Exploiting semidefinite relaxation and matrix rank minimizationtechniques, we formulate it as a semidefinite program.3. An iterative algorithm, combining the augmented Lagrange function methodand quasi-Newton method with the step-size rule of inexact line search, isshown to solve the SDP problem.