根据交通风险动态决定是否调用云端视觉模型，降低通信开销。
Risk-Adaptive Edge--Cloud Visual Reasoning for Communication-Efficient Autonomous Driving
- 通过车载系统评估交通风险，决定是否请求云端推理
- 云请求减少54.1%，自动紧急刹车次数下降，任务成功率相当
- 适合对通信效率和实时性要求高的自动驾驶场景
云端视觉语言模型（VLM）比车载小型模型具备更强的上下文推理能力，但频繁上传图像会增加通信开销，并引入网络与推理延迟，影响战术决策。本文提出一种风险自适应的边缘-云端架构：车载交通评估模块判断何时请求云端推理。车载VLM与轻量检测器捕捉时序交通状态及路径相关危险，用于保守本地响应和选择性访问云端。云端提供战术建议，而验证、车辆控制与自动紧急制动仍由本地完成。在CARLA实验中，该方法在任务成功率与周期性云端访问相当的前提下，将云端请求减少54.1%，并显著降低自动紧急制动（AEB）激活次数。在延迟施工场景下，语义事件触发请求早于下次预定检查。在三种模拟网络条件下，该方法持续降低云端流量，尽管变道时间略长于周期性访问。结果表明，车载交通评估可作为选择性VLM推理的有效触发机制。
原文摘要 · Abstract (English)
Cloud-hosted vision-language models (VLMs) offer greater contextual reasoning capabilities than smaller onboard models, but frequent visual uploads increase communication overhead and add network and inference latency to tactical decisions. We present a risk-adaptive edge-cloud architecture in which onboard traffic assessment determines when cloud reasoning is requested. An onboard VLM and a lightweight detector capture temporal traffic conditions and path-relative hazards for conservative local response and selective cloud access. The cloud model provides tactical advice, while validation, vehicle control, and automatic emergency braking remain local. In CARLA experiments, our method matched the task success rate of periodic cloud access while reducing cloud requests by 54.1% and recording fewer automatic emergency braking (AEB) activations. In a delayed-roadwork ablation, semantic events triggered requests before the next scheduled audit. Across three emulated network profiles, the method continued to reduce cloud traffic, although lane changes took longer than with periodic access. Onboard traffic assessment therefore served as a practical trigger for selective VLM inference in these experiments.
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。