An Outlier Recognition Method Based on Improved CUSUM for GPS Time Series
In order to effectively identify the abnormal data in the GPS (Global Positioning System) monitoring data, the method of the CUSUM (Cumulative Sum) median control chart was introduced. Aiming at the problem that the traditional mean control graph cannot accurately identify the outliers in the actual sample data, a GPS anomaly data recognition algorithm based on the CUSUM median control chart was proposed, and the basic principles and calculation steps were given. On the basis, considering the influence of non-normal data in the calculation of the algorithm, a method of converting to normal data was given. Finally, the feasibility and effectiveness of the proposed method were verified by simulation data. The experimental results show that the proposed algorithm has a good effect. Compared with the traditional CUSUM control chart, the abnormal value recognition ability was improved, and the false alarm rate was also effectively controlled.
KeywordsControl chart CUSUM CUSUM median GPS time series Outlier recognition
The project was financially supported by the National Natural Science Foundation of China (no. 41404004).
- 1.Jiang, W., Liu, J.: Analysis of long-term deformation of reservoir using continuous GPS observations. Acta Geodaetica Cartogr. Sin. 41, 682–689 (2012). (in Chinese)Google Scholar
- 2.Xu, S., Zhang, H., Yang, Z., Wang, Z.: GPS Measuring principle and application. Wuhan University Press, Wuhan (2001). (in Chinese)Google Scholar
- 3.Jiang, W., Liu, H., Liu, W., He, Y.: CORS development for Xilongchi dam deformation monitoring. Geomatics Inf. Sci. Wuhan Univ. 37, 949–952 (2012). (in Chinese)Google Scholar
- 4.Kuang, C., Dai, W.: Measurement of wind-induced vibration of tall buildings using GPS and wavelet application. Geomatics Inf. Sci. Wuhan Univ. 35, 1024–1028 (2010). (in Chinese)Google Scholar
- 5.Kuang, C., Yi, Z., Dai, W., Zeng, F.: Measuring wind-induced response characteristics of tall building based on GPS PPP method. J. Central S. Univ. 44, 4588–4596 (2013). (in Chinese)Google Scholar
- 6.Yi, T., Guo, Q., Li, H., Lin, Y.: Multi-step identification algorithm based on relational negative selection for GPS abnormal monitoring data. J. Vib. Eng. 28, 1–8 (2015). (in Chinese)Google Scholar
- 8.Tie, J.: Statistical Methods of Quality Management. China Machine Press, Beijing (2006)Google Scholar
- 9.Yi, T., Guo, Q., Li, H.: The research on detection methods of GPS abnormal monitoring data based on control chart. Eng. Mech. 30, 133–141 (2013). (in Chinese)Google Scholar
- 15.Tran, K.P., Castagliola, P., Celano, G.: Monitoring the ratio of population means of a bivariate normal distribution using CUSUM type control charts. Statistical Papers, pp. 1–27 (2018)Google Scholar
- 16.Hu, X., Zhou, X., Huang, W., Jiang, G.: Design of the CUSUM Median Chart. J. Appl. Stat. Manage 37, 469–477 (2018). (in Chinese)Google Scholar
- 19.Xie, Z.: MATLAB statistical analysis and application: 40 case studies. Beihang University Press, Beijing (2010). (in Chinese)Google Scholar