Experimental Dataset for FDIA Detection in an M3C Platform
收藏资源简介：
1. Overview This dataset was constructed for FDIA detection in a M3C experimental platform. The data were collected under normal and attacked operating conditions and were used to train and evaluate the proposed detection model in the associated paper. 2. Data Source The dataset was generated on an experimental platform based on a typical M3C topology. All FDIA samples were injected into the information transmission links among the subcontrollers. 3. Attack Types The dataset includes the following operating cases:- No attack- Step FDIA- Ramp FDIA- Random FDIA- Stealthy FDIA 4. Input Features The input features include:1. KF-estimated capacitor voltage2. Integral of KF-estimated capacitor voltage3. Measured capacitor voltage4. Integral of measured capacitor voltage5. Reference capacitor voltageDetailed descriptions of all variables are provided in variables_description 5. Dataset Size The dataset contains 30,800 data points in total:- 10,800 normal data points- 20,000 attacked data points, including: - 5,000 step FDIA samples - 5,000 ramp FDIA samples - 5,000 random FDIA samples - 5,000 Stealthy FDIA samples 6. Notes The trained model reported in the paper was directly used for online testing. The remaining experimental settings are described in the associated manuscript. 7. Contact For questions regarding the dataset, please contact:Tingjun PanZhejiang Universitypan_tingjun@zju.edu.cn
1. 概述 本数据集专为M3C实验平台下的虚假数据注入攻击（False Data Injection Attack，以下简称FDIA）检测任务构建。数据集采集自正常运行与受攻击运行工况，用于训练并评估相关论文中提出的检测模型。 2. 数据来源 本数据集生成于基于典型M3C拓扑的实验平台。所有FDIA样本均被注入至子控制器间的信息传输链路中。 3. 攻击类型 本数据集包含以下运行工况： - 无攻击工况 - 阶跃型FDIA样本 - 斜坡型FDIA样本 - 随机型FDIA样本 - 隐蔽型FDIA样本 4. 输入特征 输入特征包括： 1. 卡尔曼滤波（Kalman Filter，以下简称KF）估计的电容电压 2. KF估计电容电压的积分值 3. 实测电容电压 4. 实测电容电压的积分值 5. 参考电容电压 所有变量的详细说明请参见variables_description。 5. 数据集规模 本数据集共包含30800条数据： - 10800条正常工况数据 - 20000条受攻击工况数据，其中包含： - 5000条阶跃型FDIA样本 - 5000条斜坡型FDIA样本 - 5000条随机型FDIA样本 - 5000条隐蔽型FDIA样本 6. 备注 论文中报道的训练完成模型可直接用于在线测试。其余实验设置详见相关研究手稿。 7. 联系方式 若对本数据集有任何疑问，请联系： 潘廷军 浙江大学 pan_tingjun@zju.edu.cn
