A new method for direction finding based on Markov random field model

A new method for direction finding based on Markov random field model
复制标题

DOI:
10.1002/2014rs005635
复制
发表时间:
2015-07
期刊:
影响因子:
1.6
通讯作者:
Mamoru Ota;Y. Kasahara;Y. Goto
Mamoru Ota;Y. Kasahara;Y. Goto
中科院分区:
计算机科学4区
文献类型:
--
作者:
Mamoru Ota;Y. Kasahara;Y. Goto

文献摘要

相似文献

研究科学卫星在地球等离子体磁层/磁层中观测到的等离子体波的特征,对于了解波的产生机制和影响波的产生和传播的等离子体环境是有效的。特别是,找到波的传播方向对于理解VLF/ELF波的机制非常重要。为了找到这些方向,波分布函数(WDF)方法已被提出。这种方法是基于这样的思想,即观测信号由定义波能量密度分布的多个基本平面波组成。然而,由此产生的方程构成了一个不适定问题,其中的解不是唯一确定的;因此,必须为解假设一个适当的模型。虽然已经提出了许多模型,我们必须选择最佳的模型为给定的情况下,因为每个模型都有自己的优点和缺点。在本研究中,我们提出了一种新的方法来确定等离子体波接收机测量的等离子体波的方向。我们的方法是基于这样的假设,即WDF可以表示为一个马尔可夫随机场模型与推理的模型参数使用变分贝叶斯学习算法。使用计算机生成的谱矩阵,我们评估了模型的性能,并将结果与两种传统方法的结果进行了比较。
Investigating the characteristics of plasma waves observed by scientific satellites in the Earth's plasmasphere/magnetosphere is effective for understanding the mechanisms for generating waves and the plasma environment that influences wave generation and propagation. In particular, finding the propagation directions of waves is important for understanding mechanisms of VLF/ELF waves. To find these directions, the wave distribution function (WDF) method has been proposed. This method is based on the idea that observed signals consist of a number of elementary plane waves that define wave energy density distribution. However, the resulting equations constitute an ill‐posed problem in which a solution is not determined uniquely; hence, an adequate model must be assumed for a solution. Although many models have been proposed, we have to select the most optimum model for the given situation because each model has its own advantages and disadvantages. In the present study, we propose a new method for direction finding of the plasma waves measured by plasma wave receivers. Our method is based on the assumption that the WDF can be represented by a Markov random field model with inference of model parameters performed using a variational Bayesian learning algorithm. Using computer‐generated spectral matrices, we evaluated the performance of the model and compared the results with those obtained from two conventional methods.