A location-aware solution for predicting driver’s destination in intelligent traffic systems

Abstract

This paper presents a methodology focused on learning driving tendencies using GPS data and time stamps to forecast future movement locations. People move throughout regions of time in established, but variable patterns and a person’s normal movement can be learned by machines. Location extraction from raw GPS data in combination with a probabilistic neural network is proposed for learning human movement patterns. Using time as an input over a distribution of data, normal tendencies of movement can be forecasted by analyzing the probabilities of a target being at a specific point within a set of frequented locations. This model can be used to predict future traffic conditions and estimate the effects of using the same routes each day. Considering traffic density on its own is insufficient for a deep understanding of the underlying traffic dynamics and hence we propose a novel method for automatically predicting the capacity of each road segment. We evaluate our method on a database of GPS routes and demonstrate their performance. Ultimately, the results produced can contribute to the prediction of traffic congestion in urban areas.