arXiv 2506.085412025-07-05cs.CVcs.AI

TrajFlow：基于流匹配的多模态运动预测

TrajFlow: Multi-modal Motion Prediction via Flow Matching

University of British Columbia（不列颠哥伦比亚大学）；Vector Institute for AI（向量人工智能研究所）；University of Waterloo（滑铁卢大学）；Georgia Institute of Technology（佐治亚理工学院）；Carnegie Mellon University（卡内基梅隆大学）；Tesla（特斯拉）；XPeng Motors（小鹏汽车）；Leapmotor；Nvidia（英伟达）；DiDi（滴滴出行）；Canada CIFAR AI Chair（加拿大 CIFAR 人工智能主席）

Qi Yan, Brian Zhang, Yutong Zhang, Daniel Yang, Joshua White, Di Chen, Jiachao Liu, Langechuan Liu, Binnan Zhuang, Shaoshuai Shi, Renjie Liao

备注：: IROS 2025

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

TrajFlow提出基于流匹配的多模态运动预测框架，通过单次推理生成多条轨迹，结合Plackett-Luce排序损失和自条件训练，在WOMD上取得最先进性能。

摘要：

高效且准确的运动预测对于确保自动驾驶中的安全性和知情决策至关重要，尤其是在需要多模态预测的动态现实世界条件下。我们提出了TrajFlow，一种新颖的基于流匹配的运动预测框架，旨在解决现有生成式轨迹预测方法在可扩展性和效率方面的挑战。与采用独立同分布采样并需要多次推理来捕获多样结果的常规生成方法不同，TrajFlow在单次推理中预测多条合理的未来轨迹，显著降低了计算开销，同时保持了预测之间的一致性。此外，我们提出了一种基于Plackett-Luce分布的排序损失，以改进预测轨迹的不确定性估计。我们还设计了一种自条件训练技术，在第二次前向传播中重用模型自身的预测来构造含噪输入，从而提高泛化能力并加速推理。在大规模Waymo Open Motion Dataset（WOMD）上的大量实验表明，TrajFlow在多个关键指标上取得了最先进的性能，凸显了其在安全关键的自动驾驶应用中的有效性。代码和其他细节可在项目网站https://traj-flow.github.io/上获取。

英文摘要：

Efficient and accurate motion prediction is crucial for ensuring safety and informed decision-making in autonomous driving, particularly under dynamic real-world conditions that necessitate multi-modal forecasts. We introduce TrajFlow, a novel flow matching-based motion prediction framework that addresses the scalability and efficiency challenges of existing generative trajectory prediction methods. Unlike conventional generative approaches that employ i.i.d. sampling and require multiple inference passes to capture diverse outcomes, TrajFlow predicts multiple plausible future trajectories in a single pass, significantly reducing computational overhead while maintaining coherence across predictions. Moreover, we propose a ranking loss based on the Plackett-Luce distribution to improve uncertainty estimation of predicted trajectories. Additionally, we design a self-conditioning training technique that reuses the model's own predictions to construct noisy inputs during a second forward pass, thereby improving generalization and accelerating inference. Extensive experiments on the large-scale Waymo Open Motion Dataset (WOMD) demonstrate that TrajFlow achieves state-of-the-art performance across various key metrics, underscoring its effectiveness for safety-critical autonomous driving applications. The code and other details are available on the project website https://traj-flow.github.io/.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。