GPS based Smartphone Users' Personal Route Model Learning and Moving Route Prediction

Nowadays, much research has focused on managing personal locations using GPS sensors for more intelligent LBSs. In this paper, we propose efficient learning and prediction methods for personal routes. Firstly, we have researched learning method for routes using image processing algorithms. This approach has shown reduction of complexities and enhancement of performance such as accuracy and speed. Secondly, we propose a method for prediction on a user's paths using a small segment of current GPS logs and the user's learned routes. The accuracy of route prediction is 94%, when we evaluated the proposed methods which had been implemented on smartphones.