摘要
针对GPS载波相位观测值小周跳探测与修复的困难,提出了一种利用改进的集总平均经验模态分解进行单频载波相位周跳探测、并用最小二乘支持向量机算法预测来修复周跳的方法。首先利用相位观测值及伪距观测值构造周跳探测量;其次对探测量进行改进的集总平均经验模态分解,得到多个本征模式分量,取相关系数较高的IMF分量进行Hilbert谱分析,用Hilbert谱中模极大值的位置来判断周跳发生位置;最后对相关系数较高的IMF分量利用最小二乘支持向量机方法进行预测,比对发生周跳历元的载波实测值与预测值以修复周跳。实验表明,此方法可成功探测并修复一周的小周跳。
To solve the problem that it is difficult to detect and repair the small cycle-slip, a method which uses modified EEMD to detect the cycle-slip and uses LSSVM to repair the cycle-slip is proposed. First, pseudorange and carrier-phase measurements are used to construct the cycle-slip detectable quantity;Second, the quantity is decomposed by MEEMD;as a result, it can obtain some intrinsic mode functions (IMF). Choose the IMFs which possessed relatively high correlativity with the detectable quantity to do the analysis use Hilbert. And the Hilbert modulus maximum is applied to judge if the cycle-slip appeared and its epoch. Finally, the value at these epochs for every IMFs whose correlation is high is predicted by the LSSVM. Compare the predictive value with the real value and the cycle-slip can be repaired. The experiments show that this method can successfully detect and repair the cycle-slip as small as one cycle.
作者
朱江淼
闫迪
高源
陈烨
王星
ZHU Jiang-miao;YAN Di;GAO Yuan;CHEN Ye;WANG Xing(Beijing University of Technology Faculty of Information Technology, Beijing 100124, China;National Institute of Metrology, Beijing 100029, China)
出处
《计量学报》
CSCD
北大核心
2019年第3期491-497,共7页
Acta Metrologica Sinica
基金
国家自然科学基金(2013YQ090943
20140010)
