Detection and Estimation of Erroneous Positioning Data
Ha Yoon Song, Kyoung-Woon On · Procedia Computer Science · 2013
Improvement of mobile devices lead functionality of positioning and many of related applications using positioning data are widely used. However the positioning data acquired by such devices are prone to have erroneous values due to environmental factors. In this research, a detection and correction algorithm of erroneous data over continuous positioning data set. Our algorithm is based on moving window of speed derived by consecutive positioning data. Moving average and moving standard deviation in moving window will compose a moving significant interval at a given time, and the moving significant interval is utilized in order to detect the erroneous positioning data along with other parameters. Not only estimates our algorithm the correct position, but also removes the effect of errors to moving window statistics. Estimation of possible speed is firstly done and then estimation of possible position is also done. Experimental verification based on our algorithm will be also presented.