HDGT: Heterogeneous Driving Graph Transformer for Multi-Agent Trajectory Prediction via Scene Encoding.
Xiaosong JiaPenghao WuLi ChenHongyang LiYu LiuJunchi YanPublished in: CoRR (2022)
Keyphrases
- multi agent
- prediction accuracy
- prediction model
- d scene
- image sequences
- single image
- video sequences
- three dimensional
- location prediction
- graph representation
- reinforcement learning
- directed graph
- real scenes
- scene understanding
- prediction error
- multiple images
- graph structure
- moving points
- weighted graph
- random walk
- graph matching
- fault diagnosis
- multi agent systems
- inter frame
- bipartite graph
- scene analysis
- video scene
- inter layer
- structured data