GPS-simulated trajectory detection

Abstract

Due to the prevalence of GPS-enabled devices and wireless communication technology, spatial trajectories have become the basis of many location based applications, e.g., Didi. However, trajectory data suffers low quality problems causing by sensor errors and artificial forgeries. Sensor errors are inevitable while forgeries are always constructed on bad purpose. For example, some Didi drivers use GPS simulators to generate forgery trajectories and make fake transactions. In this work we aim to distinguish whether a given trajectory is a GPS simulated trajectory. By formulating this task as the problem of traffic speed extracting and irregular measuring, we propose a simulated trajectory detection framework. In traffic speed extracting phase, we first divide time into time slots and then extract the regular speed of each road during each time slot. In irregular measuring phase, we propose three methods to measure the distance between the speed of the given trajectory and the real traffic speeds. For empirical study, we apply our solution to a real trajectory dataset and have found that the simulated trajectory detection framework can detect most forgery trajectories.