Compressive sensing based estimation of OFDM nonlinear distortion

Abstract

Compressive sensing (CS) has recently been proposed to recover nonlinearly distorted orthogonal frequency division multiplexing (OFDM) signals. This approach is only effective when the nonlinear distortion is very sparse, and as a result the power amplifier (PA) needs to have a large linear range. However, practical PAs typically have a small linear range to improve power efficiency and reduce cost. In this case, the nonlinear distortion has poor sparsity, which results in large estimation errors with CS-based methods. Further, the out-of-band radiation caused by nonlinear PA distortion cannot be mitigated since CS-based techniques are performed at the receiver. To address these problems, phase rotated pilot signals are introduced to modify the OFDM signal at the transmitter to ensure that the clipping noise is sufficiently sparse. This improves the performance of CS-based methods for PAs with a small linear range. In addition, the proposed technique reduces the peak-to-average power ratio (PAPR) and thus also the out-of-band radiation.