[论文] Edge-Assisted Crowsensing for Explainable Cyclist Safety Perception

论文概要

研究领域: CV
作者: Hongsen Cao, Chia-Yen Chiang, Mona Jaber, Meihui Wang, Yasmin Fathy
发布时间: 2026-11-05
arXiv: 2611.08441

中文摘要

尽管政策鼓励主动出行，骑行者死亡率仍然居高不下，自动驾驶车辆也给混合交通增加了复杂性。事故统计和多模态检测流水线都无法扩展到路线级别的连续安全感知监测。我们提出一个边缘辅助的众包感知系统：车把摄像头提供街景帧，视觉-语言模型（VLM）通过单一 prompt 将每帧转化为结构化交通属性，可解释提升机（EBM）将这些属性与 OpenStreetMap 道路基础设施和车载运动学数据结合，预测众包的 1-9 安全评分。提取理应放在网络边缘：最快的自托管模型每帧需 2.49 秒，而评分在 CPU 上仅需 0.4 毫秒。在来自伦敦两条路线的 1,362 帧标注数据上，EBM 回归器在五折交叉验证下达到 72.2% 二值准确率和 1.011 的平均绝对误差。与黑盒集成模型的差距在 1 个百分点以内。其逐特征 shape 函数提供精确归因：物理车道隔离值 +0.97 分，划线车道 +0.70 分，施工场地 -0.65 分。规划者因此可以用评分点数来定价基础设施改造，结果可直接进入成本效益分析。我们还报告了标注者间一致性（alpha = 0.4695）、属性审计、以及属性误差与精度关联的敏感性分析。通过重编码，20 秒采样间隔可为 1,000 名骑行者服务，带宽 54 Mbit/s，精度代价 1.4 个百分点，使连续监测切实可行。

原文摘要

Cyclist fatality rates remain high despite policy support for active travel, and autonomous vehicles add complexity to mixed traffic. Neither crash statistics nor multi-modal detection pipelines scale to continuous monitoring of perceived safety along a route. We present an edge-assisted crowdsensing system: handlebar cameras supply street-level frames, a vision-language model (VLM) turns each frame into structured traffic attributes through a single prompt, and an explainable boosting machine (EBM) combines them with OpenStreetMap road infrastructure and on-bike kinematics to predict a crowdsourced 1-9 safety score. Extraction belongs at the network edge: the fastest self-hosted model needs 2.49 s per frame and scoring takes 0.4 ms on a CPU. On 1,362 annotated frames from two London routes, the EBM regressor reaches 72.2 percent binary accuracy and a mean absolute error of 1.011 under five-fold cross-validation. It is within one accuracy point of black-box ensembles. Its per-feature shape functions give exact attribution: physical lane separation is worth +0.97 rating points, a painted lane +0.70 and a construction site -0.65. Planners can therefore price an infrastructure change in rating points, and the results enter cost-benefit analysis directly. We also report inter-annotator agreement (alpha = 0.4695), an attribute audit, and a sensitivity analysis linking attribute error to accuracy. With re-encoding, a 20 s sampling interval serves 1,000 riders at 54 Mbit/s at a cost of 1.4 accuracy points, making continuous monitoring practical.

自动采集于 2026-11-08

#论文 #arXiv #CV #小凯