Research On The PAPR Reduction Schemes Based On Clipping And Nonlinear Distortion Recovery In OFDM Systems

Orthogonal frequency division multiplexing(OFDM)has been widely adopted for use in many wireless communication systems including DAB,WiMAX,4G LTE and IEEE 802.11,due to its robustness against multipath fading,high bandwidth efficiency and simple implementation.However,high peak-to-average power ratio(PAPR)of the OFDM signal would drive high power amplifier(HPA)at the transmitter into saturation,producing in-band distortion that degrades the bit error rate(BER)performance and outof-band distortion that corrupts the spectrum of the signals.To avoid nonlinear distortion,high PAPR requires a HPA with a large linear dynamic range or a relatively large input backoff(IBO)at the cost of power efficiency.PAPR reduction techniques can be broadly classified into three main categories: signal distortion techniques,multiple signaling and probabilistic techniques,and coding techniques.Of these techniques,clipping is the simplest signal distortion method.However,clipping is a nonlinear process that causes both in-band and out-of-band distortions.The former can degrade the BER performance and the latter one causes spectral spreading.Filtering the clipped OFDM signal can eliminate the out-of-band distortion and improve the BER performance,but it will lead to peak regrowth.Therefore,iterative clipping and filtering(ICF)has been proposed to both remove the out-of-band interference and suppress the regrowth of the peak power.The clipping distortion caused by clipping OFDM signal has sparsity in time domain.According to compressed sensing(CS)theory of sparse signal processing,a sparse signal can be reconstructed from its compressed observations.The main work of this paper is as follows:(1)A modified clipping noise recovery method based on CS at the receiver has been proposed.At the Nyquist sampling rate,considering the nonlinear distortion introduced by clipping and HPA as a whole,the entire nonlinear process is modeled to calculate the total distortion utilizing the similar sparsity as the clipping noise in the time domain.In the case of oversampling,the analysis of non-linear distortions based on the normal HPA model no longer applies.Therefore,based on the HPA polynomial model,the final nonlinear distortion is analyzed and deduced by using the third-order HPA polynomial model,combined with clipping process and Bussgang's theory.At the receiver,CS recovery algorithm is still adopted to restore the nonlinear distortion effectively.(2)This paper proposes an improved frequency-domain iterative scaling factor reduction algorithm based on the structure of the first-order SCF algorithm,because the traditional ICF needs multiple iterations to converge on the expected PAPR threshold and the convergence speed is slow.For each iteration,the frequency-domain noise scaling factor is selected,which has been determined in advance.And the PAPR will converge to the desired threshold in four iterations.Furthermore,the BER performance of the system is further deteriorated due to the iterative clipping distortion.In this paper,we propose an improved iterative clipping distortion recovery method combined with the noise enhancement factor based on CS,and also consider the total distortion recovery in the presence of HPA nonlinearity.Compared with the BER performance of the traditional ICF system,the improved method can effectively improve the BER performance of the system utilizing the noise enhancement factor to quantize the total nonlinear distortion.