基于重要性采样的最大似然时延估计算法
针对多径环境中的时延估计,提出了一种基于重要性采样概念的算法。该算法利用蒙特卡罗算法（MC）对未知参数的分布函数抽样,获得简化似然函数的全局最优解,进而通过计算样本均值直接得到参数估计结果。该方法避免了耗时较长的多维网格搜索和对初值较为敏感的迭代算法,且能够无条件收敛至全局最优值。仿真结果表明,在相同样本条件下,该算法相比于EM、MUSIC算法,不仅消除了对初值的依赖性,也获得了更接近克拉美罗界（CRLB）的仿真结果。将该算法与其他多种算法进行计算复杂度分析后发现,IS-based算法较其他算法更为简单,计算量更低,具有较为重要的工程应用价值。...
Saved in:
| Published in | 实验室研究与探索 Vol. 34; no. 12; pp. 16 - 20 |
|---|---|
| Main Author | |
| Format | Journal Article |
| Language | Chinese |
| Published |
山东服装职业学院,山东泰安,271000%黄淮学院信息工程学院,河南驻马店,463000
2015
|
| Subjects | |
| Online Access | Get full text |
| ISSN | 1006-7167 |
Cover
| Summary: | 针对多径环境中的时延估计,提出了一种基于重要性采样概念的算法。该算法利用蒙特卡罗算法（MC）对未知参数的分布函数抽样,获得简化似然函数的全局最优解,进而通过计算样本均值直接得到参数估计结果。该方法避免了耗时较长的多维网格搜索和对初值较为敏感的迭代算法,且能够无条件收敛至全局最优值。仿真结果表明,在相同样本条件下,该算法相比于EM、MUSIC算法,不仅消除了对初值的依赖性,也获得了更接近克拉美罗界（CRLB）的仿真结果。将该算法与其他多种算法进行计算复杂度分析后发现,IS-based算法较其他算法更为简单,计算量更低,具有较为重要的工程应用价值。 |
|---|---|
| Bibliography: | CHEN Jian , TIAN Li-fang （ 1. Shandong Vocational Institute of Clothing Technology, Taian 271000, China; 2. School of Information Engineering, Huanghuai University, Zhumadian 463000, China） time-delay estimation; maximum likelihood（ML）; multipath environments; importance sampling（IS） To estimate the time delay in multipath environments,a new algorithm was proposed based on importance sampling（ IS）. This algorithm applied Monte Carlo method into the sampling of unknown parameters. The global maximum of the compressed likelihood function could be found,thus,the parameters were approximated by the average of samples. Therefore,the estimator proposed avoided the complex multidimensional grid search and iterative methods which depended seriously on the initial guess. The algorithm guaranteed convergence to the global maximum. Results of simulation showed that,this algorithm held better performance than EM and MUSIC algorithms,under the conditions of the same samples. In addition,the algorithm also showed superiority |
| ISSN: | 1006-7167 |