基于期望最大化加速算法的正交频分复用信道估计

针对基于期望最大化（EM）迭代算法的正更频分多路复用（OFDM）信道估计方法复杂度高、收敛慢，严重制约OFDM的传输速率的缺陷，提出了一种基于EM加速算法的OFDM信道估计方法．该方法基于拟牛顿加速算法，并结合一种带调整参数的Broyden对称秩1校正公式来实现，具有二次收敛性，提高了EM的计算速度，降低了计算复杂度．仿真结果表明，相比于空间选择期望最大算法，在性能损失只有0．1dB的情况下，EM算法的复杂度得到很大的降低，且其计算速度可提高十几倍．

Orthogonal Frequency Division Multiplexing Channel Estimation Method Based on Accelerate Expectation Maximization Algorithm

Aiming at the deficiencies of high complexity,slow convergence and hindering the transmission rate etc.in the orthogonal frequency division multiplexing(OFDM) channel estimation method based on expectation maximization(EM) algorithm,an acceleration EM OFDM channel estimation algorithm was proposed,which is complemented by quasi-Newton accelerate method and combines with Broyden symmetry rank one correction formulae with adjusted parameters.The algorithm has quadratic convergence and can speed up the EM computation.Simulation results show that in the case of only 0.1 dB performance loss,computing speed of the EM algorithm can be increased by more than 10 times compared to the space-alternating generalized expectation-maximization algorithm.