摘要
针对GPS／INS组合导航中广泛采用的扩展卡尔曼滤波进行数据融合难以建立精确的数学模型问题，提出采用平移不变量的97提升小波算法对观测信号进行降噪。平移不变量可有效地消除信号在不连续点产生的伪吉布斯现象，97提升算法比传统的Mallat小波更具准确性和快速性。与传统的软、硬阈值函数相比，改进的阈值函数对组合导航信号处理具有更好的连续性。研究结果表明，改进的97提升平移小波消噪方法是一种节约计算时间、提高导航性能的优越方法。
In connection with difficulty in establishing accurate mathematical model by extensively applied Extended Kalman Filter （EKF） in data process in GPS/INS Integrated Navigation, 97 lifting wavelet algorithm that using Translationinvariant method is applied to reduce noise of observational signal. Pseudo-Gibbs appearances of signal that are produced in discontinuity points can effectively be eliminated by Translation-invariant. 97 Lifting wavelet algorithm is more accurate and quicker than traditional Mallat wavelet method. Compared with traditional soft and hard threshold functions, the improved threshold function is more continuous in navigation signal process. Study shows the improved Wavelet De-noising Method is a superior method in saving calculation time and improving Navigation performance.
出处
《计量学报》
CSCD
北大核心
2010年第2期150-153,共4页
Acta Metrologica Sinica
