arXiv 2603.241552026-09-21cs.RO

面向目标的反应式仿真用于闭环轨迹预测

Goal-Oriented Reactive Simulation for Closed-Loop Trajectory Prediction

University of Wuppertal（伍珀塔尔大学）

Harsh Yadav, Tobias Meisen

备注：: The core claim that the simulation becomes reactive does not hold, even though the findings on its effects on the ego vehicle remain valid. This claim was re-evaluated based on reviewer feedback, and the paper is therefore being retracted. However, the valid results regarding the effects on the ego vehicle are being merged into the parallel study: https://arxiv.org/abs/2603.23393

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

本文提出一种闭环训练范式，通过引入基于Transformer的场景解码器，提升高频率重规划下的碰撞避免能力，实验显示在nuScenes和密集DeepScenario交叉口的碰撞率降低显著。

摘要：

当前轨迹预测模型主要以开环方式训练，在现实闭环设置中常导致协变量偏移和累积误差。此外，依赖静态数据集或非反应式日志回放模拟器切断交互循环，阻碍了自动驾驶agent主动协商周围交通的能力。本文提出一种面向目标的闭环训练范式，优化高频率递推 horizon 自身预测。通过引入基于Transformer的场景解码器，使自身预测 grounded 在现实交通互动中，实现反应式一致性。通过让agent接触混合的开环数据和模拟的自诱导状态，模型学习恢复行为以纠正自身执行错误。大量评估表明，闭环训练在高重规划频率下显著提升碰撞避免能力，在nuScenes和密集DeepScenario交叉口相比开环基线，碰撞率降低达27.0%和79.5%。此外，我们还展示了一种混合模拟，结合反应式与非反应式周围agent，实现了即时交互性与长期行为稳定性的最佳平衡。

英文摘要：

Current trajectory prediction models are primarily trained in an open-loop manner, which often leads to covariate shift and compounding errors when deployed in real-world, closed-loop settings. Furthermore, relying on static datasets or non-reactive log-replay simulators severs the interactive loop, preventing the ego agent from learning to actively negotiate surrounding traffic. In this work, we propose an on-policy closed-loop training paradigm optimized for high-frequency, receding horizon ego prediction. To ground the ego prediction in a realistic representation of traffic interactions and to achieve reactive consistency, we introduce a goal-oriented, transformer-based scene decoder, resulting in an inherently reactive training simulation. By exposing the ego agent to a mixture of open-loop data and simulated, self-induced states, the model learns recovery behaviors to correct its own execution errors. Extensive evaluation demonstrates that closed-loop training significantly enhances collision avoidance capabilities at high replanning frequencies, yielding relative collision rate reductions of up to 27.0% on nuScenes and 79.5% in dense DeepScenario intersections compared to open-loop baselines. Additionally, we show that a hybrid simulation combining reactive with non-reactive surrounding agents achieves optimal balance between immediate interactivity and long-term behavioral stability.

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。