子带OFDMA中EM双选信道估计的收敛性能分析

Convergence analysis of EM-based doubly selective channel estimation methods for subband OFDMA

算法实时性是通信系统可实用化的关键,为了揭示子带正交频分多址(OFDMA)系统中期望最大化(EM)双选信道估计的实时性,研究了子带OFDMA系统中EM算法的收敛性能.通过系统仿真方法分别给出了单天线系统和多天线系统EM迭代信道估计的收敛曲线,研究结果表明,两种迭代信道估计都有较快的收敛速度,而且多天线系统比单天线系统收敛更快,前者仅需3次迭代即可收敛,而后者在10次迭代内也可保证收敛.理论分析显示,迭代过程中参数更新数目的不同是导致两种算法收敛性能不同的原因所在.

Real-time computation of an algorithm in communications system is the key for the system to be applicable. In order to reveal whether the channel estimations based on expectation maximization （EM） algorithm in subband OFDMA systems can be on line realized or not, the convergence performance of the EM- based doubly selective channel estimation methods for subband OFDMA is investigated in the article. First, the convergence curves of the two iterative estimators, respectively, for the single transmit antenna system and the multiple transmit antennas system are stated. Results obtained from the convergence curves show that both estimators converge fast, and moreover, the latter does faster with only 3 iterations than the former with a little bigger number of about 10 iterations. Theoretical analysis indicates that the differences in the updated parameter number in iterations account for the two estimators＇ differences in convergence speed.