Nonlinear Channel Estimation for Minimally Sampled Baseband OFDM Systems

Abstract

The orthogonal frequency-division multiplexing (OFDM) is a widely used modulation scheme in modern digital communication systems. An OFDM signal can be susceptible to channel nonlinearities due to its high peak-to-average power ratio (PAR). This makes estimation and compensation of nonlinear channels particularly vital for OFDM systems. In this paper, we consider identifying cubically nonlinear channels in baseband OFDM systems using minimally sampled input and output signals. Unlike nonlinear channels in bandpass OFDM systems, which can be modeled by a complex baseband equivalent Volterra series with only odd order terms, the nonlinear channel in a baseband OFDM system requires a real Volterra series with both even and odd order terms. The minimally sampled input and output signals of the baseband OFDM system are used to estimate the frequency-domain Volterra kernels of the nonlinear channel. Due to the possible spectral regrowth caused by the channel nonlinearities, the minimally sampled output signal could be aliased. Nevertheless, by conducting higher-order statistical analyses on the OFDM signals, we show that closed form expressions of the Volterra kernels can still be derived. As a result of the closed form expressions, a simple estimate for the Volterra kernels is attained. Furthermore, the derived Volterra kernel estimate is shown to achieve the minimum mean square error (MMSE). Therefore, the proposed approach indeed provides an efficient yet accurate way to estimate nonlinear channels in baseband OFDM systems.