Driver Identification Data Processing Repository

Overview

This repository contains the code used to generate the Driver Identification Dataset. The dataset is created by processing raw data files collected from 50 participants driving a 2014 Kenworth T270 Class 6 truck. The code processes the data into three different formats:

1. High-resolution SQLite database
2. High-resolution CSV files
3. Downsampled CSV files (includes additional variables and survey answers)
4. KML file generation (for GPS visualization in Google Earth)

Repository Structure

The repository is organized into four main folders, each containing the necessary scripts to generate the corresponding data files:

1. CSVHighResolutionCode/

• Contains scripts to generate high-resolution CSV files from raw data.
• Requires all raw data files for processing.
• Contains README with more info

2. SQLiteHighResolutionCode/

• Contains scripts to generate a high-resolution SQLite database.
• Requires all raw data files for processing.
• Contains README with more info

3. DownsampledCode/

• Contains scripts to generate downsampled CSV files.
• The downsampled data includes additional computed variables (e.g., distance traveled, cyberattack active) and survey responses.
• Contains README with more info

4. KMLGeneratorCode/

• Contains the Python script to generate a KML file for GPS visualization.
• The script reads longitude and latitude from the sqlite database.
• Outputs a KML file that can be opened with Google Maps, Earth, etc.
• Script file: GenerateKLMFile.py

KML File Generation

To generate a KML file from the SQLite database, use the script GenerateKLMFile.py. This script reads the SparkfunGPSData table and extracts the longitude and latitude values.

Usage

python3 GenerateKLMFile.py

The script will create a file named DriverID#G#TruckCape.kml in the same directory. Open this file with Google Maps, Earth, etc to visualize the GPS data and route of the drives.

Ex Image from DriverIDG1S01:

Required Raw Data Files

To generate the processed datasets, the following raw data files are required for each participant:

• DriverIDSXGYCANLogger3.bin (CAN data from CANLogger3 (https://github.com/SystemsCyber/CAN-Logger-3))
• DriverIDSXGYEDA.csv (Electrodermal activity data from Empatica E4 wristband)
• DriverIDSXGYHR.csv (Heart rate data from Empatica E4 wristband)
• DriverIDSXGYIBI.csv (Interbeat interval data from Empatica E4 wristband)
• DriverIDSXGYTruckCape.log (CAN data from TruckCape Beaglebone (https://github.com/SystemsCyber/TruckCapeProjects))
• DriverIDSXGYVBOX.vbo (GPS and inertia data from VBOX 3i)

where:

• X = Subject Number (1-50)
• Y = Group Number (G1, G2, G3)

Participant Groups

• Group 1 (G1): 17 subjects
• Group 2 (G2): 16 subjects
• Group 3 (G3): 17 subjects

CAN Data Decoding Requirement

The J1939 digital annex is required to decode the CAN data. It must be converted into a JSON format following the instructions in this repository:

• pretty_j1939 Repository