SOQPSK-TG信号的EM半盲载波频偏估计

针对成形偏移四相相移键控-TG(Shaped Offset Quadrature Phase Shift Keying-Telemetry Group version, SOQPSK-TG)信号在训练序列长度受限时频偏估计精度较低的问题，利用双二进制分解(Doubinary Decomposition, DBD)原理提出基于期望最大化(Expectation Maximization, EM)的SOQPSK半盲载波频偏(Carrier Frequency Offset, CFO)估计算法。为了确保EM算法收敛到预期性能范围，使用基于非线性四次方码元定时估计算法的非数据辅助频偏估计方法优化了EM算法初始点选择。仿真实验结果表明，该算法相比于使用训练序列进行数据辅助估计的方法，在不增加辅助数据数量的前提下能够进一步提升CFO估计的精度，并在较高信噪比下拥有接近序列总长度所对应的克拉美罗界(Cramér-Rao Bound, CRB)的优秀性能。

Carrier Frequency Offset Semi-blind Estimation for SOQPSK-TG Signals Based on EM Algorithm

To address the problem of low accuracy of carrier frequency offset estimation for shaped offset quadrature phase shift keying-telemetry group version(SOQPSK-TG) signals when the length of the training sequence is limited,a semi-blind carrier frequency offset(CFO) estimation algorithm based on expectation maximization(EM) and doubinary decomposition(DBD) for SOQPSK signals is proposed.In addition,to ensure that the proposed EM algorithm converges to the expected performance range,the initial point selection of the EM algorithm is optimized using a non-data-assisted frequency offset estimation method based on the nonlinear quartic symbol timing estimation algorithm.The simulation results show that the proposed algorithm can further improve the accuracy of CFO estimation without increasing the number of auxiliary data compared with the data-aided estimation method using training sequences,and it has an excellent performance close to the Cramér-Rao bound(CRB) corresponding to the total length of the sequence at a higher signal-to-noise ratio.