Comparison of accuracy and precision of GPS-enabled mobile devices

Abstract

GPS-enabled mobile devices such as smart phones and tablets are extremely popular today. Billions of such devices are currently in use. The application and research potentials of these devices are limitless, but how accurate are these devices? The research team used Average Euclidean Error (AEE), Root Mean Square Error (RMSE), and Central Error (CE) to define and calculate the accuracy and precision of six popular GPS-enabled mobile devices running on two different operating systems (Android and iOS) from measurements in Clarksville, Tennessee, USA. Spatial data (multiple measurements) consisting of latitude and longitude coordinates of positions reported by these six GPS-enabled mobile devices were collected at nine different physical locations (control points). The results of the field data were ranked and compared through the use of one scientific and two realistic measuring protocols. The scientific protocol was designed to simulate standard scientific experimental practices (a controlled experiment). The realistic protocols were designed to simulate daily usage of mobile devices (holding a mobile phone at around waist level). In this preliminary study, it was discovered that various mobile devices perform differently in terms of AEE, RMSE, and CE. It also showed that the protocol is not a statistically significant contributing factor to the variation in the measurements in terms of accuracy nor precision. The device and the location are statistically significant contributing factors to the variation in the measurements in terms of accuracy and precision. The interaction between protocol and device is not a statistically significant contributing factor. It is also discovered that the devices are statistically less accurate than they are precise.

Publication
2016 IEEE International Conference on Computer and Information Technology (CIT)