一种基于迭代的多音干扰消除方案

摘要

For multi-tone interference in orthogonal frequency division multiplexing (OFDM) system,one scheme is proposed,which first increases frequency domain resolution of multi-tone interference identification,and then performs parameter estimation,reconstruction and elimination on several single-tone interference one by one based on iterative computation.In multi-tone interference identification stage,ZoomFFT technology is used to improve the frequency domain resolution and simplify the computational complexity.However,down sampling sequence number after ZoomFFT does not meet the n power of 2.For single-tone interference parameters estimation,the classical maximum likelihood parameter estimation based on DFT (DFTMLE) can be used.But when the number of sampling points is too large,DFT operation will bring great implementation complexity.Based on DFT-MLE,this paper presents an FFT based single-tone parameters estimation FFT-MLE.It not only reduces the computational complexity,but also provides better robustness on estimation performance compared with DFT-MLE.The simulation results show that this multi-tone interference cancellation scheme is effective.%针对正交频分复用技术(OFDM)系统的多音干扰,本文提出先通过增加频域分辨率进行多音干扰识别,再基于迭代运算逐个对多个单音干扰进行参数估计、重构和消除的方案.其中,在多音干扰识别阶段,使用ZoomFFT技术以提高频域分辨率并简化计算复杂度.然而,ZoomFFT后的降采样序列点数不一定满足2的整数次幂,对于单音干扰参数估计,此时可采用经典的基于离散傅里叶变换(DFT)的最大似然参数估计(DFT-MLE),但当采样点数太大时,DFT运算将带来很大的实现复杂度.本文在基于DFT-MLE算法基础上给出一种基于快速傅里叶变换(FFT)的单音参数估计法FFT-MLE,不仅简化了运算复杂度,而且相比DFT-MLE其估计性能有更好的鲁棒性.最后的仿真结果说明了该多音干扰消除方案的有效性.