Self-driving vehicles rely on multimodal motion forecasts to effectively interact with their environment and plan safe maneuvers. We introduce SceneMotion, an attention-based model for forecasting scene-wide motion modes of multiple traffic agents. Our model transforms local agent-centric em-beddings into scene-wide forecasts using a novel latent context module. This module learns a scene-wide latent space from multiple agent-centric embeddings, enabling joint forecasting and interaction modeling. The competitive performance in the Waymo Open Interaction Prediction Challenge demonstrates the effectiveness of our approach. Moreover, we cluster future waypoints in time and space to quantify the interaction be-tween agents. We merge all modes and analyze each mode independently to determine which clusters are resolved through interaction or result in conflict. Our implementation is available at: https://github.com/kit-mrt/future-motion
SceneMotion: From Agent-Centric Embeddings to Scene-Wide Forecasts
24.09.2024
2520577 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
BEHAVIOR PREDICTION USING SCENE-CENTRIC REPRESENTATIONS
Europäisches Patentamt | 2025
|Scene-Centric Identification and Retrieval of Unmanned Aerial Vehicle (UAV) Video Segments
British Library Conference Proceedings | 2000
|Online Contents | 1996