A Novel Statistical Model of OFDM Signals with Optimal Tone Reservation PAPR Reduction for EVM Theoretical Analysis

摘要

In this paper, we study the statistical distribution of the OFDM time-domain signal amplitude to which the optimal tone reservation (TR) PAPR reduction solution is applied. This study comes as a major step towards the derivation of the analytical expression of the error vector magnitude (EVM) metric on OFDM waveforms when non-linear amplification is used. We demonstrate through simulations that this distribution has a bimodal behavior and can be tightly modeled as a superposition of two probability density functions whose expressions are given and whose parameters are obtained through proper estimation. From such analysis, we provide the complete distribution expression of the PAPR-reduced OFDM signal, which can be further exploited for the derivation of performance bounds of the TR PAPR reduction strategy.