Predicting a Pedestrian Trajectory Using Sequence to Sequence Learning for Mobile Robot Navigation

Abstract

This paper proposes a method to predict the future trajectory of a pedestrian as sequence data by using massive trajectory records collected by various sensor devices. We aim to use the method for safely and efficient path planning of autonomous mobile robots in a human-robot coexisting environment. For the prediction, we use Sequence to Sequence learning, which is frequently used in the field of natural language processing. That enables to treat long-term sequence data. In order to verify the effectiveness of the proposed method, we conduct experiments using a dataset. As a result, we show that we can predict the trajectory sufficiently by converting the trajectory data to adequate sequence data.