BiTraP: Bi-directional Pedestrian Trajectory Prediction with Multi-modal Goal Estimation
Abstract Pedestrian trajectory prediction is an essential task in robotic applications such as autonomous driving and robot navigation. State-of-the-art trajectory predictors use a conditional variational autoencoder (CVAE) with recurrent neural networks (RNNs) to encode observed trajectories and decode multi-modal future trajectories. This process can suffer from accumulated errors over long prediction horizons (>=2 seconds). This …
Read more “BiTraP: Bi-directional Pedestrian Trajectory Prediction with Multi-modal Goal Estimation”