一种非平稳非线性频谱占用度时间序列分析方法

A Method for Analysis of Non-linear and Non-stationary Spectrum Occupancy Time Series

针对传统频谱占用度分析模型由于未考虑序列的非线性非平稳特性,导致无法准确描述频谱占用度特性的问题,该文提出将集合经验模式分解(EEMD)方法与人工神经网络(ANN)的方法结合应用于频谱占用度时间序列建模方法中,采用EEMD+ANN的频谱占用度序列建模和预测方法.首先应用EEMD分解算法把原始频谱占用度时间序列分解成不同尺度的基本模态分量,再根据不同尺度的基本模态分量分别构建ANN模型,提高了模型针对复杂频谱占用度时间序列的学习能力.结合实测数据分析,表明该模型相对传统频谱占用度模型具有更高的拟合和预测精度,验证了该方法的正确性与有效性.

In order to analyze the non-linear and non-stationary spectrum occupancy time series which cannot be directly analyzed base on traditional time series method,a novel prediction modelling method of spectrum occupancy time series based on ensemble empirical mode decomposition（ EEMD） and Artificial Neural Network（ ANN） is proposed. Firstly,the spectrum occupancy time series is decomposed into serval intrinsic model function（ IMF） so as to make every component stationary. Then in viewof the stationary time series,a prediction ANN model is established correspondingly for each IMF. Simulative experiment for practical measured data shows that the proposed method has higher precision in comparison with other methods,i. e.,effective to non-linear and ono-stationary complicated spectrum occupancy time series prediction.