Simplified ICF and smart MIR for PAPR reduction in OFDM systems

Abstract

One of the biggest problems for orthogonal frequency division multiplexing (OFDM) systems is the high peak-to-average power ratio (PAPR), which breaks the orthogonality among subcarriers and leads to the nonlinear distortion of transmitted signals after being processed by the power amplifier (PA). The iterative clipping and filtering (ICF) method is one of the well known and applied existing PAPR reduction techniques at the transmitter and the modified iterative receiver (MIR) is an effective existing method for signal recovery with the ICF method in its iterative process at the receiver. However, the ICF method, as well as the MIR, suffers from the high computational complexity due to the oversampling and high-order inverse fast Fourier transform/fast Fourier transform (IFFT/FFT) operators. Besides, the performance of MIR is limited by the number of iterations. In this paper, to reduce the computational complexity of ICF method, the phase rotation iterative clipping and filtering (PRICF) method is proposed, which performs padding, phase rotation and low-order IFFT/FFT operators. Meanwhile, the computational complexity of MIR is also reduced because the ICF method is replaced by the PRICF method in its iterative process. Furthermore, to accelerate the iteration or improve the performance, the modified iterative network receiver (MIR-Net) is proposed by introducing trainable parameters based on the method of model-driven deep learning. Comparing with the combination of ICF and MIR, the simulation results demonstrate the advantages of our proposed methods, which is the combination of PRICF and MIR-Net, in terms of the computational complexity and performance.