Pedestrian trajectory prediction with convolutional neural networks

• Published in: Pattern recognition Vol. 121; p. 108252
• Main Authors: Zamboni, Simone, Kefato, Zekarias Tilahun, Girdzijauskas, Sarunas, Norén, Christoffer, Dal Col, Laura
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.01.2022
• Subjects: Autonomous driving; Convolution; Convolutional neural network; Convolutional neural networks; Crowd surveillance; Data driven modelling; Forecasting; Gaussian noise (electronic); Model-based OPC; Neural-networks; Pedestrian prediction; Pedestrian trajectories; Physics-based modeling; Recurrent neural networks; Trajectories; Trajectory prediction; Convolutional neural networks; Trajectory prediction; Pedestrian prediction
• ISSN: 0031-3203; 1873-5142; 1873-5142
• DOI: 10.1016/j.patcog.2021.108252

•New convolutional model achieves state-of-the-art results on ETH and TrajNet datasets.•Random rotations and Gaussian noise are the best data augmentation techniques.•Coordinates with the origin in the last observation point better represent trajectory. Predicting the future trajectories of pedestrians is a challenging problem that has a range of application, from crowd surveillance to autonomous driving. In literature, methods to approach pedestrian trajectory prediction have evolved, transitioning from physics-based models to data-driven models based on recurrent neural networks. In this work, we propose a new approach to pedestrian trajectory prediction, with the introduction of a novel 2D convolutional model. This new model outperforms recurrent models, and it achieves state-of-the-art results on the ETH and TrajNet datasets. We also present an effective system to represent pedestrian positions and powerful data augmentation techniques, such as the addition of Gaussian noise and the use of random rotations, which can be applied to any model. As an additional exploratory analysis, we present experimental results on the inclusion of occupancy methods to model social information, which empirically show that these methods are ineffective in capturing social interaction.