研究人员开发了DeCoR，一个新颖的强化学习框架，旨在优化城市街道设计和交通信号控制。该系统首先通过将行人网络编码为图来学习生成最优的过街设施布局。随后，它开发自适应信号配时，以最大限度地减少行人和车辆的延误。在真实城市走廊的模拟中，DeCoR显著减少了行人等待时间并改善了交通流量，证明了其对不同需求和布局变化的鲁棒性。 AI
影响 这项研究可能带来更高效的城市规划和交通管理系统，减少拥堵并提高行人安全。
排序理由 该集群包含一篇详细介绍新研究框架及其模拟结果的学术论文。
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
arXiv:2605.21311v1 Announce Type: cross Abstract: Modern vision systems can detect, track, and forecast urban actors at scale, yet translating perception outputs to urban design remains limited. We introduce DeCoR, a two-stage reinforcement learning framework that leverages flow …
Modern vision systems can detect, track, and forecast urban actors at scale, yet translating perception outputs to urban design remains limited. We introduce DeCoR, a two-stage reinforcement learning framework that leverages flow observations to co-optimize crosswalk layout and n…
Modern vision systems can detect, track, and forecast urban actors at scale, yet translating perception outputs to urban design remains limited. We introduce DeCoR, a two-stage reinforcement learning framework that leverages flow observations to co-optimize crosswalk layout and n…