TAGCN: Topology-Aware Graph Convolutional Network for Trajectory Prediction
Document Type
Conference Proceeding
Publication Date
12-8-2020
Publication Title
International Symposium on Visual Computing
First page number:
542
Last page number:
553
Abstract
Predicting future trajectories of agents in a dynamic environment is essential for natural and safe decision making of autonomous agents. The trajectory of an agent in such an environment not only depends on the past motion of that agent but also depends on its interaction with other agents present in that environment. To capture the effect of other agents on trajectory prediction of an agent, we propose a three stream topology-aware graph convolutional network (TAGCN) for interaction message passing between the agents. In addition, temporal encoding of local- and global-level topological features are fused to better characterize dynamic interactions between participants over time. Results are competitive compared to previous best methods for trajectory prediction on ETH and UCY datasets and highlights the need for both local and global interaction structure.
Keywords
Trajectory prediction; Graph embedding; Graph convolutional network
Disciplines
Computer Sciences | Numerical Analysis and Scientific Computing | Physical Sciences and Mathematics
Language
English
Repository Citation
Biswas, A.,
Morris, B. T.
(2020).
TAGCN: Topology-Aware Graph Convolutional Network for Trajectory Prediction.
International Symposium on Visual Computing
542-553.
http://dx.doi.org/10.1007/978-3-030-64556-4_42