An Adaptive Learning-Based Approach for Vehicle Mobility Prediction

This work presents an innovative methodology to predict the future trajectories of vehicles when its current and previous locations are known. We propose algorithm adapt trajectories’ data based on consecutive GPS construct a statistical inference module that can be used online for mobility prediction. The is hidden Markov model (HMM), where each trajectory modeled as subset locations. prediction stage uses information inferred so far Viterbi algorithm, which identifies (hidden information) with maximum likelihood prior known (observations). By analyzing disadvantages using (TDVIT) number states increases, we enhanced (OPTVIT), decreases computation time. Offline analysis vehicle conducted through evaluation dataset containing real traces 442 taxis running in city Porto, Portugal, during full year. Experimental results obtained show process improved more about available. Moreover, time significantly OPTVIT adopted approximately 90% performance achieved, showing effectiveness proposed method