Abstract: Security is one of the important factors in every trip, traditional methods are mostly based on human or video collaboration to complete security monitoring, and there is no way to measure travel by trajectory data. With the technological improvement of positioning and tracking systems, the continuous development and popularization of GPS satellite positioning technology makes it possible to collect and use GPS data points on car navigation devices and handheld GPS devices.This paper proposes a method for evaluating security based on user GPS trajectory data, and uses trajectory data between dormitory and laboratory in the campus as the research object to verify the effectiveness of the method. Research shows that the user trajectory off the safety trajectory obtains a lower safety score. Conversely, if the user trajectory is very similar to the safety trajectory, a high safety score is obtained. This method is very efficient in verifying whether the user trajectory is similar to the safety trajectory or not, and at the same time, the form of the trajectory data reflects the user’s travel safety.

Key words: GPS trajectory, travel safety, K-means clustering, similarity of Euclidean distance, distance safety score