Predicting Future Locations for Moving Objects

In the context of transportation optimization, the possibility to predict future movements for moving object is a crucial innovation in order to make the best decision in terms of time, cost, and impact on the environment. Unfortunately, future location prediction is a challenging task. Existing works exploit techniques to predict a trip destination, but such proposals are effective only when location data are precise (e.g., collected with GPS tracking systems), and movements are observed for long periods of time (e.g., weeks). In this work we introduce an approach based on a Hidden Markov Model which overcomes these limits and improves existing results in terms of precision of the prediction, for both the final destination and the route (i.e., trajectory) to get there. This new model is resistant to uncertain location data, as it works with data collected by using cell-towers to localize the users instead of GPS devices, and reaches good prediction results in much shorter times (days instead of weeks in a representative application). Finally, in order to enable the use of the technology in an application for the Social Web, we introduce an enhanced version of the model with a speed-up in the execution times measurable in order of magnitudes w.r.t. the standard HMM implementation. To show the applicability of the approach we show an application of the technology in the context of a carpooling service. Carpooling is an effective way to reduce travel costs, but the increasing complexity of work and social schedules have made it less desirable and have kept rather low the level of participation. In particular, drivers want flexibility in their schedule and the possibility to commit to others only when they are certain about their time and route. We show how our model enables a novel “real-time” carpooling, where drivers do not have to commit to potential passengers and do not have to manually define the start and end points for their trips.