Non-redundant Peak-to-average Power Ratio Reduction For F-OFDM Systems

Filtered-orthogonal frequency division multiplexing(F-OFDM)is a promising new radio candidate waveform for the fifth generation wireless communications.It has been considered as a potential alternative waveform to OFDM for the 5th generation mobile communication(5G)heterogeneous networks,with the following advantages.Firstly,FOFDM system divides the assigned bandwidth into several unequal-sized sub-bands,and different communication needs are accommodated in different sub-bands with the most appropriate waveform parameters.F-OFDM can reduce the out-of-band emission(OOBE)by using a finite impulse response(FIR)filter in each sub-band and save the bandwidth of guard band with the affordable system complexity.Thirdly,F-OFDM can use the non-contiguous spectral to improve the spectral efficiency(SE).We research on reducing the high peak-to-average-power ratio(PAPR)of F-OFDM signal at the transmitter.The PAPR reduction algorithms for F-OFDM sub-band signals and sub-band splicing signals are presented in this thesis.In this thesis,we propose a non-redundant hybrid PAPR reduction scheme which includes linear precoding,time-domain selected mapping(TSLM)with low computational complexity and companding techniques,for F-OFDM sub-bands.The proposed three-step PAPR reduction scheme outperforms the previous conventional PAPR reduction methods.Hybrid PAPR reduction scheme consider the practical issue of power amplifier(PA)efficiency in PAPR reduction for F-OFDM systems.In light of the fact that the PA efficiency generally increases with the input signal power,we reduce the PAPR while maintaining the signal power as close to the PA saturation point as possible by companding,to enable the PA to work with the highest possible efficiency,which can reduce the costs of communication link without bit error rate(BER)performance loss.The sub-band splicing signals can increase peak power and PAPR.The crest factor reduction algorithm can control PAPR of sub-band splicing signals and avoid spectrum leakageThe reciprocal of the PAPR reduction at the receiver is embedded in the ambiguity elimination process of the independent component analysis(ICA)based blind channel equalization.A non-redundant linear precoding can introduce the correlation between the original signal and the reference data to eliminate the phase ambiguity caused by ICA and TSLM together and therefore does not require any dedicated side information from the transmitter or any exclusive signal processing at the receiver.F-OFDM system with ICA based blind equalization can not onlyreduce the complexity,but also improve the spectral efficiency.According to the simulation results,the BER performance of the system with the proposed hybrid PAPR reduction scheme is shown to be close to zero forcing performance with perfect channel state information(CSI),while no side information and training sequence are required for PAPR reduction and channel estimation,thanks to the effectiveness of the ICA based blind channel equalization.