arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2510.24166cs.AI
收藏

UniPlanner:一种面向自动驾驶车辆决策系统的多数据集融合统一运动规划框架

UniPlanner: A Unified Motion Planning Framework for Autonomous Vehicle Decision-Making Systems via Multi-Dataset Integration

Guangdong Key Laboratory of Intelligent Information Processing, College of Electronics and Information Engineering, Shenzhen University(广东智能信息处理重点实验室,电子与信息工程学院,深圳大学);School of Computer Science and Cyber Engineering, Guangzhou University(计算机科学与网络工程学院,广州大学);School of Artificial Intelligence, Guangzhou University(人工智能学院,广州大学)

Xin Yang, Yuhang Zhang, Wei Li, Xin Lin, Wenbin Zou, Chen Xu

AI总结:

针对自动驾驶运动规划单数据集训练鲁棒性不足的问题,提出首个多数据集融合规划框架UniPlanner,通过HFTDN、GFTM和S2D三项创新实现跨数据集统一学习,提升规划安全性与效率。

摘要:

运动规划是自动驾驶车辆决策系统的核心组件,直接决定轨迹安全性与行驶效率。尽管深度学习方法已提升了规划能力,但现有方法仍局限于单数据集训练,限制了规划的鲁棒性。\n 通过系统分析,我们发现车辆轨迹分布以及历史-未来关联在不同数据集间呈现出显著的一致性。基于这些发现,我们提出UniPlanner——首个专为自动驾驶决策领域多数据集融合设计的规划框架。UniPlanner通过三项协同创新实现跨数据集统一学习。\n 第一,历史-未来轨迹字典网络(History-Future Trajectory Dictionary Network, HFTDN)聚合来自多个数据集的历史-未来轨迹对,利用历史轨迹相似性检索相关未来轨迹,生成跨数据集规划引导。\n 第二,无梯度轨迹映射器(Gradient-Free Trajectory Mapper, GFTM)从多数据集中学习鲁棒的历史-未来关联,将历史轨迹转化为通用规划先验。其无梯度设计确保在引入有价值先验的同时防止捷径学习,使规划知识可安全迁移。第三,稀疏到稠密(Sparse-to-Dense, S2D)范式采用自适应dropout,在训练期间选择性抑制规划先验以实现鲁棒学习,同时在推理阶段充分利用先验以最大化规划性能。

英文摘要:

Motion planning is a critical component of autonomous vehicle decision-making systems, directly determining trajectory safety and driving efficiency. While deep learning approaches have advanced planning capabilities, existing methods remain confined to single-dataset training, limiting their robustness in planning. Through systematic analysis, we discover that vehicular trajectory distributions and history-future correlations demonstrate remarkable consistency across different datasets. Based on these findings, we propose UniPlanner, the first planning framework designed for multi-dataset integration in autonomous vehicle decision-making. UniPlanner achieves unified cross-dataset learning through three synergistic innovations. First, the History-Future Trajectory Dictionary Network (HFTDN) aggregates history-future trajectory pairs from multiple datasets, using historical trajectory similarity to retrieve relevant futures and generate cross-dataset planning guidance. Second, the Gradient-Free Trajectory Mapper (GFTM) learns robust history-future correlations from multiple datasets, transforming historical trajectories into universal planning priors. Its gradient-free design ensures the introduction of valuable priors while preventing shortcut learning, making the planning knowledge safely transferable. Third, the Sparse-to-Dense (S2D) paradigm implements adaptive dropout to selectively suppress planning priors during training for robust learning, while enabling full prior utilization during inference to maximize planning performance.

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。

↑