论文/arXiv 2610.13781·19 天前

SimVTA：面向驾驶世界动作模型的仿真支撑视频-轨迹对齐

SimVTA用仿真渲染匹配视频补足候选轨迹监督，并在NAVSIM与HUGSIM上验证驾驶WAM后训练。

AI 整理

• 本体 SDK 与控制
• 仿真与训练
• 具身模型

原文链接：SimVTA: Simulation-Grounded Video-Trajectory Alignment for Driving World Action Models（arXiv 2610.13781 · 发布日期：2026-10-09）

SimVTA: Simulation-Grounded Video-Trajectory Alignment for Driving World Action Models

SimVTA用仿真渲染匹配视频补足候选轨迹监督，并在NAVSIM与HUGSIM上验证驾驶WAM后训练。

核心要点

• 驾驶日志只含记录轨迹的未来视频，替代轨迹缺少视觉监督。
• 框架采样候选轨迹，并在3DGS重建场景中渲染高奖励未来视频。
• 在有限渲染预算下优先监督高误差未来，并用GRPO与流匹配联合优化。
• NAVSIM v1上PDMS由86.2升至91.3，HUGSIM闭环HD-Score为38.8。

验证范围与局限：验证范围为NAVSIM v1/v2和HUGSIM闭环评估。

英文摘要 (Abstract)

Driving world action models (WAMs) learn from paired trajectories and future observations. Post-training with reinforcement learning allows the policy to explore alternative trajectories and improve planning using driving rewards. However, driving logs contain future videos only for the recorded trajectories, leaving these alternatives without corresponding visual supervision. Thus, we introduce SimVTA, a simulation-grounded post-training framework that combines trajectory optimization with video supervision. First, we sample candidate trajectories for each observation and use a real-world simulator (e.g., World Engine) to render future videos for high-reward candidates in real-world scenes reconstructed with 3D Gaussian Splatting (3DGS). This provides trajectory-matched supervision independent of the model's own predictions. Under a limited rendering budget, we prioritize video supervision for higher-error futures among these rendered candidates. Finally, we optimize trajectory planning with Group Relative Policy Optimization (GRPO) and use a flow-matching loss on the matched videos to update shared representations. At inference time, the model predicts trajectories directly, without simulator calls. On NAVSIM v1, applying SimVTA to a WAM initialization improves PDMS from 86.2 to 91.3 (+5.1). It achieves 88.2 EPDMS on NAVSIM v2 and an average HD-Score of 38.8 in closed-loop evaluation on HUGSIM. Ablations show that updating shared representations with video supervision improves planning, while prioritizing higher-error futures among high-reward trajectories further improves performance.

作者：Changhui Jing, Haochen Tian, Hongchen Li, Tianyu Li, Caojun Wang, Zihao Zhang, Yanhao Liu, Chengen Xie, Bailin Li, Hao Zhao

来源

本页事实对应的公开页面，以原文为准。

1. arXiv 2610.13781：SimVTA: Simulation-Grounded Video-Trajectory Alignment for Driving World Action Models访问 2026-10-21