A fast real-time algorithm for outlier elimination based on polynomial modeling
Zhangming He, Tao Ren, Chen Zhu, Juhui Wei, Guosheng Li · 2019 IEEE 8th Data Driven Control and Learning Systems Conference (DDCLS) · 2019
Data recorded by the recording system often appear outliers. In order to correctly reflect and predict the falling process of aircraft debris, it is necessary to deal with the outliers. In this paper, MATLAB/MATHEMATICA symbolic math toolbox is used to solve the inverse matrix of the design matrix, which eases the reasoning by our brain, because the formula is very complicated. Based on the result above, the polynomial fitting and recursive formula are deduced for filtering and prediction, and finally the theoretical result are successfully applied for outlier detection and reconstruction of the aircraft target debris trajectory. The results show that this method seems quite feasible and can eliminate the outliers effectively.