HAI 1.0: HIL-based Augmented ICS Security Dataset
Talkat 13th USENIX Workshop on Cyber Security Experimentation and Test (CSET '20) · on Monday, 2020-08-10·
summary
HAI 1.0 is a hardware-in-the-loop ICS dataset built from a GE turbine, Emerson boiler and FESTO water treatment system synchronized via a dSPACE HIL simulator, with automated setpoint changes to run long benign traces and scalable automated attacks. Attack tooling implements 14 primitives including response prevention and setpoint manipulation to create 38 attack scenarios; datasets include two training runs (≈7 days and ≈4 days) and testing runs (28 and 10 attacks), 63 columns per record, and are published on GitHub and Kaggle.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 0:54What is the ICS Security Dataset
- 1:54Problem Solving
- 2:54Process Augmentation
- 3:54Test Data
- 4:31Tech Tool
- 5:11Attack Primitive
- 6:13Attack Implementation
- 7:35Results
- 8:35Feature
- 9:02Contest
related talks
14mAutoHack: A Physically Verified Multi-Bus CAN Dataset for Intrusion Detection...
32mIntroduction of Cybersecurity AI Dataset
40mMe, My Adversary & AI: Investigating & Hunting with Machine Learning
10mOpenSynth Dataset Extension - 7K Challenge
55mAttacking And Protecting Artificial Intelligence
50mAI Systems Are Software Systems
40mMaking Sense of the Landscape of Attacks and Defenses Against AI
32mI need data (for research)!