摘要
北斗卫星导航系统(BDS)卫星钟差异常值处理过程中,由于成片异常值的存在,往往会产生掩盖与淹没现象,致使异常值探测效率不高甚至失败。基于求和自回归移动平均模型,分析了时间序列中成片加性AO(Additive Outlier)异常值探测时易产生掩盖与淹没现象的原因;考虑了差分及逆差分对异常值探测的影响,提出了成片AO类异常值探测的抗掩盖与淹没新算法。通过仿真算例,验证了新算法对于序列中成片AO类异常值探测的准确性和有效性。将算法应用于BDS卫星钟差异常值探测和钟差预报中,较好地克服了数据中掩盖和淹没现象产生的影响,对于进一步提高卫星钟差的预报精度具有重要作用。
In the process of dealing with the outliers of the BDS satellite clock offset,due to the existence of the patch outliers,the phenomenon of the outliers are masked and the normal data are swamped often occurs,which leads to the low efficiency or even failure of outliers detection.Based on the ARIMA model,the reason why the additive outliers are masked and the normal data are swamped when the patch outliers are detected in time series was analyzed,the influence of difference on outlier detection was taken into account,and a new algorithm for detecting the outlier patches who can prevent the masking and swamping was proposed.Simulation examples illustrate that the proposed algorithm for detecting the patch additive outliers is accurate and effective.When it is applied in the outlier detection and forecasting of the BDS satellite clock offset,the occurrence of masking and swamping can be prevented efficiently,the outliers can be detected and the magnitudes of the additive outliers accurately can be estimated in the clock offset sequence.Furthermore,it can obtain higher prediction precision when this new algorithm is applied to process the data of the BDS satellite clock offset.
作者
马朝忠
张倩倩
韩松辉
李新娜
MA Chaozhong;ZHANG Qianqian;HAN Songhui;LI Xinna(Information Engineering University,Zhengzhou 450001,China;31002 Troops,Beijing 100094,China)
出处
《测绘科学技术学报》
北大核心
2019年第4期346-352,共7页
Journal of Geomatics Science and Technology
基金
国家自然科学基金项目(41474009
41174005)
