CubeSat cybersecurity dataset for intrusion detection (CuCD-ID): Labelled NOS3/cFS telemetry (raw + augmented) with COSMOS reproduction scripts

用于入侵检测的立方体卫星网络安全数据集(CuCD-ID):带有 COSMOS 复现脚本的已标记 NOS3/cFS 遥测数据(原始数据 + 增强数据)

阅读:3

Abstract

This data article presents the CubeSat Cybersecurity Dataset for Intrusion Detection (CuCD-ID), a collection of labelled command and telemetry data designed to support machine learning-based security research for space systems. The data were generated in a high-fidelity software-in-the-loop environment using NASA's Operational Simulator for Small Satellites (NOS3) running the core Flight System (cFS). Telemetry was captured across five scripted scenarios: one nominal case and four adversarial tactics aligned with the Space Attack Research and Tactic Analysis (SPARTA) framework, specifically command flooding, false data injection, storage exhaustion, and defence impairment. All scenarios were driven by commands issued from the COSMOS v4 ground station software. The repository contains two primary tabular datasets in Comma-Separated Values (CSV) format: a raw, balanced dataset with 25,000 records and 31 features, and an augmented, noised dataset with 22,465 records and 23 features. Each record contains features parsed from Consultative Committee for Space Data Systems (CCSDS) packet headers or engineered from a 20-second sliding window, alongside system-level metrics and a numeric class label. The augmented data incorporates nine documented noise categories to emulate plausible in-orbit disturbances and improve model robustness against benign variability: White Noise, Analog Outliers, Gaps, Trends, Signal Shifts, Frequency Changes, Sensor Dropout, Magnitude Warping, and Window Time Warping. The dataset is suitable for developing and benchmarking supervised and unsupervised intrusion detection methods, including on-board and Tiny Machine Learning applications. All COSMOS v4 scripts used to generate the scenarios are also provided to ensure full reproducibility and enable extension of the data collection.

特别声明

1、本页面内容包含部分的内容是基于公开信息的合理引用;引用内容仅为补充信息,不代表本站立场。

2、若认为本页面引用内容涉及侵权,请及时与本站联系,我们将第一时间处理。

3、其他媒体/个人如需使用本页面原创内容,需注明“来源:[生知库]”并获得授权;使用引用内容的,需自行联系原作者获得许可。

4、投稿及合作请联系:info@biocloudy.com。