SPADE
收藏资源简介:
SPADE是由加拿大皇家军事学院创建的多模态仿真数据集,专为智能网联汽车中信号相位与定时(SPaT)消息的攻击检测而设计。该数据集包含约1,890,000条带标签的时间步记录,每条记录融合SPaT消息字段、车载摄像头置信度分数及V2V协同数据,共40维特征,覆盖六种攻击类型与一种良性类别。数据集通过Eclipse MOSAIC仿真环境运行时注入攻击生成,结合四种交叉口几何形状、六种操作条件及五种随机种子重复,共180个场景配置。其旨在解决车载视角下SPaT消息完整性攻击的检测难题,为深度学习入侵检测系统提供跨模态数据支撑。
SPADE is a multimodal simulation dataset developed by the Royal Military College of Canada, specifically designed for attack detection of Signal Phase and Timing (SPaT) messages in intelligent connected vehicles (ICVs). This dataset contains approximately 1,890,000 labeled time-step records, each integrating SPaT message fields, on-board camera confidence scores, and vehicle-to-vehicle (V2V) cooperative data, resulting in a total of 40-dimensional features covering six types of attacks and one benign class. The dataset is generated via runtime attack injection in the Eclipse MOSAIC simulation environment, with 180 total scenario configurations combining four intersection geometries, six operating conditions, and five replicated random seeds. It aims to address the detection challenge of SPaT message integrity attacks from the on-board vehicle perspective, providing cross-modal data support for deep learning-based intrusion detection systems.
SPADE 数据集详情
基本信息
- 数据集名称:SPADE
- 来源地址:https://github.com/jdinovo/SPADE
内容概述
该数据集详情页面仅提供了项目名称“SPADE”,未包含任何关于数据集的具体描述、构成、用途、标注信息、数据规模、格式或下载方式等实质性内容。页面中的 README 文件仅有标题,无详细说明。
补充说明
由于该页面提供的有效信息极为有限,目前无法提取数据集的具体领域、任务类型或数据特点。如需了解该数据集的详细情况,建议进一步访问其官方 GitHub 仓库或相关发布文档。




