Learning pedestrian dynamics with kriging

This paper proposes a method for learning pedestrian dynamics with kriging, which is a spatial interpolation method in geosciences. Pedestrian dynamics is generally restricted by other pedestrians and its restriction is caused by social interaction between them. In the proposed method, the social interaction is represented by spatio-temporal correlation of pedestrian dynamics and the correlation is estimated by kriging. As an application of the proposed method, the prediction of pedestrian movement is examined and its performance is evaluated with publicly available benchmark dataset. The experimental results show that 10-step ahead prediction is successful with more than 80% trajectories of the datasets if 2.0[m] distance error is allowed.