广义频分复用(Generalized Frequency Division Multiplexing,GFDM)是能够满足低轨通信卫星各种应用和场景要求的候选波形.为了在基于GFDM的低轨卫星通信系统中实现可靠的数据传输,信道估计(Channel Estimation,CE)是先决条件.但是由于GFDM的二维块调制和子载波之间的非正交性使得传统CE方法几乎不可能直接应用于GFDM.除此之外,当依靠硬判决策略来保证迭代信道估计准确度的时候,需要较多迭代次数才能达到理想效果.针对上述问题,提出了一种基于软符号的迭代信道估计(Iterative Channel Estimation,ICE).根据最大后验(Maximum A Poste-riori,MAP)准则推导出ICE,通过调和平均相结合的方法构造每一个子载波上的软符号,并利用不动点迭代(Fixed-point Iteration,FPI)求解.最后通过仿真实验证明,与硬判决信道估计相比,在同样的迭代次数下,误码率(Bit Error Rate,BER)性能提升2 dB左右,并且在低迭代次数下,该方法的均方误差(Mean Squared Error,MSE)性能明显低于硬判决迭代信道估计,更适合对运算复杂度有较高要求的低轨卫星通信系统.
Iterative Channel Estimation Based on Soft Symbols in GFDM Systems
Generalized Frequency Division Multiplexing(GFDM)is a candidate waveform that can meet the various applications and scenario requirements of low Earth orbit(LEO)communication satellites.In order to achieve reliable data transmission in LEO satellite communication systems based on GFDM,channel estimation(CE)is a prerequisite.However,due to the two-dimensional block modulation of GFDM and the non-orthogonality between subcarriers,traditional CE methods are almost impossible to apply directly to GFDM.In addition,when relying on hard decision strategies to ensure the accuracy of iterative channel estimation,a large number of iterations are required to achieve the desired effect.To address these issues,a soft symbol-based iterative channel estimation(ICE)method is proposed.The ICE is derived according to the Maximum A Posteriori(MAP)criterion,and soft symbols for each sub-carrier are constructed using a harmonic mean combination method,with solutions obtained through fixed-point iteration(FPI).Finally,simulation experiments demonstrate that compared to hard decision chan-nel estimation,the bit error rate(BER)performance improves by about 2dB under the same number of iterations,and the mean squared error(MSE)performance of this method is significantly lower than that of hard decision iterative channel estimation under a low number of iterations.This makes it more suitable for LEO satellite communication systems with high computational complexity requirements.
GFDMsoft symboliterative channel estimationmaximum a posteriori(MAP)criterionlow-orbit satellite communication system
NETL
NSTL
万方数据
