Goal-based Trajectory Prediction for improved Cross-Dataset Generalization

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Grimm, Daniel, Abouelazm, Ahmed, Zöllner, J. Marius
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916861455630336
author Grimm, Daniel
Abouelazm, Ahmed
Zöllner, J. Marius
author_facet Grimm, Daniel
Abouelazm, Ahmed
Zöllner, J. Marius
contents To achieve full autonomous driving, a good understanding of the surrounding environment is necessary. Especially predicting the future states of other traffic participants imposes a non-trivial challenge. Current SotA-models already show promising results when trained on real datasets (e.g. Argoverse2, NuScenes). Problems arise when these models are deployed to new/unseen areas. Typically, performance drops significantly, indicating that the models lack generalization. In this work, we introduce a new Graph Neural Network (GNN) that utilizes a heterogeneous graph consisting of traffic participants and vectorized road network. Latter, is used to classify goals, i.e. endpoints of the predicted trajectories, in a multi-staged approach, leading to a better generalization to unseen scenarios. We show the effectiveness of the goal selection process via cross-dataset evaluation, i.e. training on Argoverse2 and evaluating on NuScenes.
format Preprint
id arxiv_https___arxiv_org_abs_2507_18196
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Goal-based Trajectory Prediction for improved Cross-Dataset Generalization
Grimm, Daniel
Abouelazm, Ahmed
Zöllner, J. Marius
Machine Learning
To achieve full autonomous driving, a good understanding of the surrounding environment is necessary. Especially predicting the future states of other traffic participants imposes a non-trivial challenge. Current SotA-models already show promising results when trained on real datasets (e.g. Argoverse2, NuScenes). Problems arise when these models are deployed to new/unseen areas. Typically, performance drops significantly, indicating that the models lack generalization. In this work, we introduce a new Graph Neural Network (GNN) that utilizes a heterogeneous graph consisting of traffic participants and vectorized road network. Latter, is used to classify goals, i.e. endpoints of the predicted trajectories, in a multi-staged approach, leading to a better generalization to unseen scenarios. We show the effectiveness of the goal selection process via cross-dataset evaluation, i.e. training on Argoverse2 and evaluating on NuScenes.
title Goal-based Trajectory Prediction for improved Cross-Dataset Generalization
topic Machine Learning
url https://arxiv.org/abs/2507.18196