【作者】 徐冠雷； 王孝通； 徐晓刚； 朱涛； 田恒斗；

【Author】 Xu Guanlei, Wang Xiaotong, Xu Xiaogang, Zhu Tao, Tian Hengdou(Dalian Naval Academy, Dalian 116018)

【机构】 海军大连舰船学院；

【摘要】 基于经验模式分解的局域波法是非平稳非线性信号时频分析的一种新算法,它的分解基是由动态信号自身确定的,具有自调节自适应的特征,将信号从高频到低频的顺序进行逐层分解,提供时域信息和不同频率信息发生的时间信息,可得到信号的瞬时频率,其性能优于传统傅立业变换,比小波分解算法具有更强的频域分辨能力。根据局域波时频分析特点和舰船航迹低频变化的特性,本文提出一种基于经验模式分解的导航数据卡尔曼滤波融合新算法,对上一时间段数据的经验模式分解获取不同分辨率下的数据,自适应求取高频噪声数据的方差,对卡滤波中噪声方差进行实时修正。试验仿真证明了该算法的有效性。更多还原

【Abstract】 Local wave analysis based on empirical mode decomposition is one non - steady time - frenqency analysis algorithm, whose base is decided by the data and has adaptive features, which decompose the data into different - level empirical modes. This algorithm can gives the time for different frenquency data, which is better than wavelet analysis and Fourier transform. According to the features and the low frenquency characteristics, one Klaman filter is proposed, in which latest data is decomposed into intrinsic mode functions and through the first function the variance of noise is estimated. The variance of the filter is updated by the real- time variance estimated.Experiments prove the efficency of this algorithm.更多还原