CAFUC2

=============================================================================
A High-Fidelity Multi-Model Benchmark Dataset for General Aviation
=============================================================================

Authors: Jing Lu, Yali Fang, Ziyi Huang, Yuan Zhang, Songhan Fan, Xuan Wu
Affiliation: Nanjing University of Aeronautics and Astronautics (NUAA), Civil Aviation Flight University of China (CAFUC)
Contact: fyl@cafuc.edu.cn (Yali Fang)
License: Creative Commons Attribution 4.0 International (CC BY 4.0)

1. OVERVIEW
—————————————————————————–
The CAFUC2 dataset is a high-fidelity multivariate time-series benchmark 
designed for General Aviation (GA) anomaly detection. It contains approx. 
1 million data points (sampled at 1 Hz) derived from 120 real flight sorties.
The fleet includes four heterogeneous aircraft models: Cessna 172R, 
Cessna 172S, Cirrus SR20, and Cirrus SR20 G6. 

To provide robust ground-truth labels for safety-critical events, we 
applied a Physics-Based Fault Injection framework to generate four types of 
anomalies, ensuring thermodynamic lag and kinematic consistency.

2. DIRECTORY STRUCTURE
—————————————————————————–
The dataset is organized under the root directory `CAFUC2/` as follows:

CAFUC2/
├── normal_data/                   # Cleaned baseline flight logs
│   ├── C172R/                     # Categorized by aircraft model
│   ├── C172S/
│   ├── SR20/
│   └── SR20G6/           
│       └── cleaned_XXXX_filled.csv  # XXXX: 4-digit flight ID
│
├── abnormal_data/                 # Flight logs with synthetic faults
│   ├── accelerator_operation/     # Throttle Surge (RPM instability & thermal lag)
│   ├── course_deviation/          # Flight Path Deviation (Kinematic lateral drift)
│   ├── engine_power_loss/         # Cylinder Misfire (Partial Power Loss)
│   └── pitch_attitude/            # Pitch Excursion (Aerodynamic G-force coupling)
│       └── clean_data_[Model]/
│           └── X.csv              # X: Extracted integer ID corresponding to baseline
│

3. FILE CORRESP