GoIRL: Graph-Oriented Inverse Reinforcement Learning for Multimodal Trajectory Prediction

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Pei, Muleilan, Shi, Shaoshuai, Zhang, Lu, Li, Peiliang, Shen, Shaojie
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866909660625240064
author Pei, Muleilan
Shi, Shaoshuai
Zhang, Lu
Li, Peiliang
Shen, Shaojie
author_facet Pei, Muleilan
Shi, Shaoshuai
Zhang, Lu
Li, Peiliang
Shen, Shaojie
contents Trajectory prediction for surrounding agents is a challenging task in autonomous driving due to its inherent uncertainty and underlying multimodality. Unlike prevailing data-driven methods that primarily rely on supervised learning, in this paper, we introduce a novel Graph-oriented Inverse Reinforcement Learning (GoIRL) framework, which is an IRL-based predictor equipped with vectorized context representations. We develop a feature adaptor to effectively aggregate lane-graph features into grid space, enabling seamless integration with the maximum entropy IRL paradigm to infer the reward distribution and obtain the policy that can be sampled to induce multiple plausible plans. Furthermore, conditioned on the sampled plans, we implement a hierarchical parameterized trajectory generator with a refinement module to enhance prediction accuracy and a probability fusion strategy to boost prediction confidence. Extensive experimental results showcase our approach not only achieves state-of-the-art performance on the large-scale Argoverse & nuScenes motion forecasting benchmarks but also exhibits superior generalization abilities compared to existing supervised models.
format Preprint
id arxiv_https___arxiv_org_abs_2506_21121
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle GoIRL: Graph-Oriented Inverse Reinforcement Learning for Multimodal Trajectory Prediction
Pei, Muleilan
Shi, Shaoshuai
Zhang, Lu
Li, Peiliang
Shen, Shaojie
Computer Vision and Pattern Recognition
Robotics
Trajectory prediction for surrounding agents is a challenging task in autonomous driving due to its inherent uncertainty and underlying multimodality. Unlike prevailing data-driven methods that primarily rely on supervised learning, in this paper, we introduce a novel Graph-oriented Inverse Reinforcement Learning (GoIRL) framework, which is an IRL-based predictor equipped with vectorized context representations. We develop a feature adaptor to effectively aggregate lane-graph features into grid space, enabling seamless integration with the maximum entropy IRL paradigm to infer the reward distribution and obtain the policy that can be sampled to induce multiple plausible plans. Furthermore, conditioned on the sampled plans, we implement a hierarchical parameterized trajectory generator with a refinement module to enhance prediction accuracy and a probability fusion strategy to boost prediction confidence. Extensive experimental results showcase our approach not only achieves state-of-the-art performance on the large-scale Argoverse & nuScenes motion forecasting benchmarks but also exhibits superior generalization abilities compared to existing supervised models.
title GoIRL: Graph-Oriented Inverse Reinforcement Learning for Multimodal Trajectory Prediction
topic Computer Vision and Pattern Recognition
Robotics
url https://arxiv.org/abs/2506.21121