用Transformer持续生成交通场景，支持动态进出车辆的长时模拟
SceneStreamer: Continuous Scenario Generation as Next Token Group Prediction
- 将交通场景建模为令牌序列，逐步生成信号、车辆状态与运动向量
- 可在无限时长内持续引入和移除车辆，生成多样且自适应的交通行为
- 生成场景训练的强化学习模型更具鲁棒性，适合自动驾驶系统评估
真实的交互式交通模拟对自动驾驶系统的训练与评估至关重要。然而，现有数据驱动的模拟方法多依赖静态初始化或日志回放数据，难以建模动态、长时程且不断变化的车流。我们提出SceneStreamer，一种统一的自回归框架，将整个场景表示为包括交通灯信号、代理状态和运动向量在内的令牌序列，并通过Transformer模型逐步生成。该设计使SceneStreamer能在无界时长内持续引入和淘汰车辆，支持逼真的长时间模拟。实验表明，SceneStreamer生成的场景具有真实感、多样性与适应性；在生成场景中训练的强化学习策略展现出更优的鲁棒性与泛化能力，验证了其作为高保真自动驾驶仿真环境的有效性。更多信息见 https://vail-ucla.github.io/scenestreamer/。
原文摘要 · Abstract (English)
Realistic and interactive traffic simulation is essential for training and evaluating autonomous driving systems. However, most existing data-driven simulation methods rely on static initialization or log-replay data, limiting their ability to model dynamic, long-horizon scenarios with evolving agent populations. We propose SceneStreamer, a unified autoregressive framework for continuous scenario generation that represents the entire scene as a sequence of tokens, including traffic light signals, agent states, and motion vectors, and generates them step by step with a transformer model. This design enables SceneStreamer to continuously introduce and retire agents over an unbounded horizon, supporting realistic long-duration simulation. Experiments demonstrate that SceneStreamer produces realistic, diverse, and adaptive traffic behaviors. Furthermore, reinforcement learning policies trained in SceneStreamer-generated scenarios achieve superior robustness and generalization, validating its utility as a high-fidelity simulation environment for autonomous driving. More information is available at https://vail-ucla.github.io/scenestreamer/ .
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。