Research On The Method Of Reducing Peak-to-Average Power Ratio In Orthogonal Time Frequency Space

Orthogonal frequency division multiplexing(OFDM),as a multi-carrier modulation technology,has excellent performance in linear time-invariant channels by virtue of its high spectral efficiency and anti-multipath interference capabilities.However,the performance of OFDM in high-Doppler extended time-varying channels(such as high-speed railway mobile communications)will drop rapidly.When applying orthogonal time-frequency space(OTFS)in a high Doppler extended channel,each transmitted symbol can obtain a constant channel gain,which can well cope with the time delay and Doppler effect in a fast time-varying channel.OTFS is a new type of multi-carrier technologies,and the pros and cons of the peak-to-average power ratio(PAPR)is one of the focuses of current researches.In order to reduce the OTFS peak-to-average ratio,this paper proposes the threshold-selective mapping(T-SLM)method and iterative-partial transfer sequence(I-PTS)method.Based on the basic principles of OTFS and multi-carrier signal PAPR,this paper analyzes the PAPR of OTFS.The results show that the PAPR of OTFS signals is linearly related to the time division(Doppler)N,and when N is large,the PAPR causes the signal to exceed the linear range of the power amplifier,resulting in signal distortion and increased bit error rate.Therefore,based on the traditional selective mapping method and the partial transmission sequence method,this paper proposes the T-SLM method and the I-PTS method.First,the influence of factors is studied,such as the number of random phase sequence vectors,phase sequence and threshold value in the T-SLM method on the PAPR of the OTFS signals.Then,the influence of factors is simulated and analyzed,such as the division method,the number of sub-block divisions and the rotation factor in the I-PTS method on the PAPR of the OTFS signals.Finally,this paper analyzes and compares the computational complexity and system error rate of the two algorithms.The experimental results show that without reverse impacts on the system's bit error rate performance,when the complementary cumulative distribution function value is equal to10^-3,the T-SLM method proposed in this paper reduces the computational complexity when the number of random phase sequence vectors is 6,the phase sequence is {1,-1},and the threshold is 9d B.In addition,the peak-to-average ratio performance of OTFS can be improved by 2.1d B.While the I-PTS method can improve the PAPR performance of OTFS by 2.7d B when the random segmentation is {1,j} and the number of segments is 4,with rotation factor fixed.Finally,in order to show the performance of the two methods,the T-SLM method,I-PTS method,? law companding method and iterative clipping filtering are compared from three aspects,i.e.,the PAPR performance,computational complexity,and system error rate.Neither the T-SLM method nor the I-PTS method affects the system bit error rate.The I-PTS method has the best PAPR performance.While the T-SLM method achieves a good compromise between the peak-to-average ratio performance and computational complexity.