BiTraP: Bi-directional Pedestrian Trajectory Prediction with Multi-modal Goal Estimation
arXiv:2007.14558
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 paper presents BiTraP, a goal-conditioned bi-directional multi-modal trajectory prediction method based on the CVAE. BiTraP estimates the goal (end-point) of trajectories and introduces a novel bi-directional decoder to improve longer-term trajectory prediction accuracy. Extensive experiments show that BiTraP generalizes to both first-person view (FPV) and bird's-eye view (BEV) scenarios and outperforms state-of-the-art results by ~10-50%. We also show that different choices of non-parametric versus parametric target models in the CVAE directly influence the predicted multi-modal trajectory distributions. These results provide guidance on trajectory predictor design for robotic applications such as collision avoidance and navigation systems.
arXiv commentMain paper: 8 pages, 5 figures, 5 tables Supplement: 4 pages, 2 figrues, 3 tables
References in corpus (10)
- Social Force Model for Pedestrian Dynamics
- Trajectory Forecasts in Unknown Environments Conditioned on Grid-Based Plans
- Bio-LSTM: A Biomechanically Inspired Recurrent Neural Network for 3D Pedestrian Pose and Gait Prediction
- Joint Attention in Autonomous Driving (JAAD)
- DESIRE: Distant Future Prediction in Dynamic Scenes with Interacting Agents
- Recurrent Network Models for Human Dynamics
- When, Where, and What? A New Dataset for Anomaly Detection in Driving Videos
- Accurate and Diverse Sampling of Sequences based on a "Best of Many" Sample Objective
- DROGON: A Trajectory Prediction Model based on Intention-Conditioned Behavior Reasoning
- Learning Regularity in Skeleton Trajectories for Anomaly Detection in Videos