Volterra-series-based analytic technique to assess the power density spectrum of nonlinearly distorted OFDM signals

Abstract

A Volterra-series-based analytic technique is developed to assess the power density spectrum of orthogonal frequency division multiplexing (OFDM) signals distorted by frequency- independent or frequency-dependent nonlinearities. The nonlinearly distorted OFDM signal power density spectrum is expressed in terms of the non-distorted OFDM signal harmonic power density spectra and the nonlinearity transfer functions. The effectiveness of the technique in both the frequency-independent and the frequency-dependent cases is established by comparing analysis and simulation. It is shown that if the OFDM signal is modelled rigorously as a cyclostationary stochastic process, analysis agrees perfectly with simulation. Further, it is also shown that if the OFDM signal is modelled as a stationary stochastic process, the developed mathematical expressions reduce to the familiar Bedrosian-Rice expressions. The appropriateness of this approximation for many practical purposes is verified by the very close agreement obtained between the simplified analytic and simulation results.