基于信赖域的EM算法进行多径时延估计

摘要

EM(estimate⁃maximize)algorithm can change multidimensional issue into one⁃dimensional issue and make the calculation process simple when it is used to estimate multipath time delay. But when data deficiency is serious,signal⁃to⁃noise ratio is low and the received signal is weak,the convergence speed of the algorithm is slow and sometimes it can’t even con⁃verge. On the contrary,the trust region algorithm has reliability,effectiveness and even stronger global convergence. Therefore, an algorithm combining the trust region algorithm with EM algorithm is developed in this paper,which can improve convergence and reduce computation complexity. The simulation results show that this algorithm can solve the problem of the slow speed of convergence and even the problem which is unable to converge.%　　EM算法进行多径时延估计时，能将多维问题转化为一维问题，计算简单，但是当数据缺失比较多，信噪比低，接收信号比较弱时收敛速度会很慢，有时甚至无法收敛。而信赖域法具有可靠性，有效性和很强的全局收敛性。因此，提出一种将信赖域法和EM算法相结合的思想，即提高了收敛特性，又降低了计算的复杂度。并通过仿真实验分析，结果表明该方法能有效解决收敛速度慢，及无法收敛的问题。