| With the development of the computer technology and the wide application of the GPS devices, it is becoming important to record the information of the trucks’ traces. The trucks’ trajectories contain a lot of business information. It is a hot topic to mine the information from the known information. However, if there are the missing trajectories, the accurate of the outcome decreases sharply. Unfortunately, a majority of countries are suffering the problem, for example, when the truck travels in the “urban canyon” or the GPS devices are turned off on purpose, this will lead to the missing trajectories. What makes matter complicate, most of the missing trajectories are discarded or roughly estimated without a high performance algorithm, which will affect the subsequent. This paper proposes an algorithm to complete the missing trajectories based on the trucks’ trajectories, which lays a foundation to analyse the drivers’ behaviour, scheduling the logistic vehicles and so on.We have read a lot of literatures which research the trajectory to deal with GPS data. In the need of the paper, first, the GPS data should be scaled to the standard including the longitude, the latitude and the speed. Next, the outliers must be mined out and deleted.To find out the missing trajectories, this paper adopts the time interval and the distance of the two adjacent GPS points. First, it is necessary to calculate the time interval and distance of the two adjacent GPS point based on the information of the GPS data. Last, the missing trajectories can be found out based on the threshold.With the unknown information of the road networks, this paper proposes two methods to complete the missing trajectories. One is to find out all the possible trajectories based on the start point and the destination of the missing trajectory. Next, we should calculate all the possible routes’ time intervals and the distances. Last, we adopt the K-modes algorithm to divide the trajectories to different groups and complete the missing trajectory based on the result of the K-modes. The other one is to calculate the area, the time interval and the distance differences which are called the STZ algorithm. Next, we adopt the K-modes algorithm to divide the trajectories to different groups and complete the missing trajectory.With the information of the road networks, this paper matches the start points and end point of the missing trajectory to the digital map data, and then adopts the A* algorithm to search the shortest routes to complete the missing trajectory with the proper trajectory. |
