Research On Probabilistic Schemes For Peak Power Optimization In OFDM Systems

Orthogonal frequency division multiplexing(OFDM)has been applied to a variety of communication standards and systems on account of its strong anti-interference and anti-multipath fading capabilities.However,its biggish peak-to-average power ratio(PAPR)causes signal distortion,restricts its large-scale use.With the purpose of abating the PAPR value of OFDM,three types of peak power optimization methods were proposed,namely predistortion,coding and probabilistic methods.Contrasted to other kinds of methods,the probabilistic method not only has good PAPR performance,but also does not emerge signal distortion.However,it has two main disadvantages such as the large computational complexity and the requirement for side information.On this basis,this thesis focuses on the reaserch of probabilistic peak power optimization methods.Root in their flaws,three optimization methods are proposed as following.1?Aiming at avoiding the side information,a peak power optimization method based on Arnold transform without side information is proposed.This method uses the augmenting sequences to augment the binary sequence and the frequency domain data sequence successively.The purpose is to make the augmented sequences meet the requirement of Arnold transform.Then the ATM is used to perform Arnold transform on the augmented sequences twice,and all OFDM candidate signals are obtained through IFFT.Finally,the candidate signal with the smallest PAPR is selected for transmitting.Compared with the original OFDM,the proposed method has better PAPR performance.More importantly,measured up against probabilistic methods like original SLM and original PTS,the demodulation of proposed method does not require any side information.2?For the flaws of original SLM method with high computational complexity,a linear combination and value sorting SLM method is proposed.In proposed method,two sets of mutually orthogonal phase factors are used to generate two sets of original candidate signals.Then the signals in each group are combined to generate two sets of new candidate signals according to the linear nature of IFFT.Next,part of signals with the high PAPR value from the two sets of original candidate signals are combined to generate a new set of candidate signals.Finally,the minimun PAPR signal is opted from five sets of OFDM candidate signals for transmitting.Measured up against the original SLM and existing Green OFDM method,this method reduces the complexity evidently while obtaining the similar PAPR reduction capacity.3?Considering the high computational complexity of original PTS method,a grouping and conjugate weighting PTS method is proposed.This method first groups the time domain subblocks in pairs,and performs conjugate and phase weighting operations on the odd and even subblocks separately in each group.Then the original subblocks,the conjugated subblocks and the weighted subblocks are combined to generate subcandidate signals.Finally,the subcandidate signals from different groups are combined to generate all OFDM candidate signals,and the one with the smallest PAPR is opted.Compared with original PTS and the existing low-complexity PTS,this method can not gain similar or better PAPR reduction capacity,but obviously abates the complexity.Finally,this paper presents the issues should be further studied,and gives the direction of future reasearch.