Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Zhou, Hao, Qi, Lu, Li, Jason, Zhang, Jie, Liu, Yi, Yang, Xu, Fan, Mingyu, Luo, Fei
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2603.10597
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866914386236407808
author Zhou, Hao
Qi, Lu
Li, Jason
Zhang, Jie
Liu, Yi
Yang, Xu
Fan, Mingyu
Luo, Fei
author_facet Zhou, Hao
Qi, Lu
Li, Jason
Zhang, Jie
Liu, Yi
Yang, Xu
Fan, Mingyu
Luo, Fei
contents Trajectory prediction is critical for autonomous driving, enabling safe and efficient planning in dense, dynamic traffic. Most existing methods optimize prediction accuracy under fixed-length observations. However, real-world driving often yields variable-length, incomplete observations, posing a challenge to these methods. A common strategy is to directly map features from incomplete observations to those from complete ones. This one-shot mapping, however, struggles to learn accurate representations for short trajectories due to significant information gaps. To address this issue, we propose a Progressive Retrospective Framework (PRF), which gradually aligns features from incomplete observations with those from complete ones via a cascade of retrospective units. Each unit consists of a Retrospective Distillation Module (RDM) and a Retrospective Prediction Module (RPM), where RDM distills features and RPM recovers previous timesteps using the distilled features. Moreover, we propose a Rolling-Start Training Strategy (RSTS) that enhances data efficiency during PRF training. PRF is plug-and-play with existing methods. Extensive experiments on datasets Argoverse 2 and Argoverse 1 demonstrate the effectiveness of PRF. Code is available at https://github.com/zhouhao94/PRF.
format Preprint
id arxiv_https___arxiv_org_abs_2603_10597
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Recover to Predict: Progressive Retrospective Learning for Variable-Length Trajectory Prediction
Zhou, Hao
Qi, Lu
Li, Jason
Zhang, Jie
Liu, Yi
Yang, Xu
Fan, Mingyu
Luo, Fei
Robotics
Artificial Intelligence
Trajectory prediction is critical for autonomous driving, enabling safe and efficient planning in dense, dynamic traffic. Most existing methods optimize prediction accuracy under fixed-length observations. However, real-world driving often yields variable-length, incomplete observations, posing a challenge to these methods. A common strategy is to directly map features from incomplete observations to those from complete ones. This one-shot mapping, however, struggles to learn accurate representations for short trajectories due to significant information gaps. To address this issue, we propose a Progressive Retrospective Framework (PRF), which gradually aligns features from incomplete observations with those from complete ones via a cascade of retrospective units. Each unit consists of a Retrospective Distillation Module (RDM) and a Retrospective Prediction Module (RPM), where RDM distills features and RPM recovers previous timesteps using the distilled features. Moreover, we propose a Rolling-Start Training Strategy (RSTS) that enhances data efficiency during PRF training. PRF is plug-and-play with existing methods. Extensive experiments on datasets Argoverse 2 and Argoverse 1 demonstrate the effectiveness of PRF. Code is available at https://github.com/zhouhao94/PRF.
title Recover to Predict: Progressive Retrospective Learning for Variable-Length Trajectory Prediction
topic Robotics
Artificial Intelligence
url https://arxiv.org/abs/2603.10597