Estimating Arrival Time of Pedestrian Using Walking Logs

Recent advances in sensor technology, wireless communication, and hand held computing devices enable location-aware services and agents that provide appropriate information and services depending on the location of the user. To provide better quality services, it is required to know and predict accurately the behavior of the user based on the data acquired from the sensors. We propose a method to estimate the arrival time of pedestrian using walking logs obtained by the GPS and stored in a database. It can be applied for pedestrian navigation and just-in-time information retrieval with high accuracy. The method retrieves walking trajectories similar to the user’s one from the database and estimates the arrival time to her destination. In the evaluation experiment, we collected walking trajectories of pedestrians who walked the same route, and compared our method with conventional ones based on the average arrival time and the extrapolation method. As a result, the estimation error of our method was smaller than that of the conventional ones. In particular, our method shows a better estimation for a route that includes a slope where the walking pattern of pedestrians drastically change.