A Bayesian Sparse Reconstruction Framework for Mitigation of Non-Linear Effects in OFDM Systems

The mitigation of nonlinear distortion caused by power amplifiers (PA) in Orthogonal Frequency Division Multiplexing (OFDM) systems is an essential issue to enable energy efficient operation. We propose a new algorithm for receiver-based clipping estimation in OFDM systems that combines the existing Iterative Hard Thresholding method with a novel Bayesian framework to estimate clipping parameters at the receiver. We avoid the use of pilots and formulate the recovery problem solely on reliably detected sub-carriers. We also develop a new criterion for selecting these reliable carriers that takes into account the channel code. Through simulations, we show that the proposed technique outperforms the existing methods both in terms of BER and speed.