Learning to Detect the Subway Station Arrival for Mobile Users

The use of traditional positioning technologies relies on the underlying infrastructures. However, for the subway environment, such positioning systems may not be available for the positioning tasks, such as the detection of the train arrivals for the passengers in the train. An alternative way is to exploit the contextual information available in the mobile devices of subway riders. To this end, in this paper, we propose to exploit multiple contextual features extracted from the mobile devices of subway riders for precisely detecting train arrivals. Along this line, we first investigate potential contextual features which may be effective to detect train arrivals according to the observations from sensors. Furthermore, we propose to explore the maximum entropy model for training a train arrival detector by learning the correlations between the contextual features and the events of train arrivals. Finally, we perform extensive experiments on several real-world data sets. Experimental results clearly validate both the effectiveness and efficiency of the proposed approach.