A Rich-Dictionary Markov Predictor for Vehicular Trajectory Forecasting

Next-location prediction in a VANET system, where each vehicle acts as a network node, is of great importance in intelligent transport systems (ITS) as this property could have a direct and positive effect on network connectivity, traffic management and hence, improve overall ITS safety. In the last few years, the widespread use of GPS navigation systems and wireless communication technology-enabled vehicles has resulted in huge volumes of trajectory data. The task of utilizing this data employing pattern-matching techniques for next-location prediction in an efficient and accurate manner is an ongoing research problem. This paper presents the Rich-Dictionary Markov Predictor (RDM), a protocol for producing online these forecasts by using a pattern matching technique. RDM is fast, accurate and fully parameterized presenting different trade-offs as regards efficiency versus prediction accuracy. We evaluated the effectiveness of RDM via simulation and the results attest that it achieves on the average more than 35% better prediction accuracy and competitive to faster prediction times than other model independent and highly accurate prediction algorithms.