📡 DATA VEIL – Synthetic Sensor Deception Engine

A research-friendly toolkit for generating trusted vs veiled sensor streams.

🧠 Overview

Data Veil is a multi-sensor deception engine that simulates how autonomous robots, drones, and IoT devices can present:

• Trusted internal data for navigation, autonomy, and safety
• Veiled external data for logs, cloud exports, or attacker-visible channels

The result is a synthetic dual-reality boundary where the system sees truth and outsiders see distortion.

🚀 Quick Start

git clone https://github.com/IoTIVP/data-veil
cd data-veil
python -m venv .venv
.venv\Scripts\activate
pip install -r requirements.txt

Run any demo:

python run_lidar_demo.py

More demos available in the repo (depth, radar, IMU, thermal, RF, stereo, ghosting, etc).

🛰 Supported Sensors & Synthetic Profiles

Data Veil currently supports synthetic veiling for:

• Depth maps
• LiDAR ranges / point clouds
• Thermal / IR
• RF field intensities
• Radar range–Doppler
• RGB camera
• Stereo vision
• Ultrasonic rings
• IMU (gyro + acceleration)
• Temporal “ghosting” sequences
• Multi-sensor dashboards

Each demo produces:

• trusted_* (internal system view)
• veiled_* (external attacker view)
• trusted_vs_hacker.png (side-by-side panel)

🎯 Threat Model & Scope

Data Veil is a simulation-only synthetic deception engine.

It is designed to distinguish between:

• Trusted internal perception (robot uses this)
• Veiled external attacker-facing perception (attackers, logs, cloud exports)

Attacker Model

We assume attackers may access:

• logs
• cloud telemetry
• debug/mirror feeds
• exported recordings

Data Veil ensures attackers only see distorted sensor data, not the true internal environment.

What Data Veil is NOT

• Not a replacement for authentication
• Not cryptographic security
• Not firmware hardening
• Not to be deployed blindly in safety-critical production

It is a research tool, not a compliance-grade security layer.

🐍 Python API (data_veil_core)

Data Veil includes a modular high-distortion sensor veiling core:

from data_veil_core import (
    veil_depth,
    veil_lidar,
    veil_radar,
    veil_thermal,
    veil_imu,
)

Example:

import numpy as np
from data_veil_core import veil_depth

depth = np.random.rand(64, 96).astype(np.float32)
veiled = veil_depth(depth, strength=1.3)

print("trusted:", depth.min(), depth.max())
print("veiled:", veiled.min(), veiled.max())

🔧 Architecture

Sensors → Trusted Pipeline → Autonomy
              ↓
          Data Veil
              ↓
    External / Untrusted Consumers

Trusted stays internal.
Veiled leaves the system.

🟣 Realtime Viewer

For a live, animated view of trusted vs veiled depth:

python examples/realtime_depth_demo.py

Features:

• Trusted vs veiled split-screen
• Synthetic depth evolving over time
• Live profile switching:
  • 1 = light
  • 2 = privacy
  • 3 = ghost
  • 4 = chaos

Profiles are loaded from config/profiles.yaml.

🟠 Multi-Sensor + Profiles Demo

To see how different profiles affect multiple sensors in one shot:

python examples/multi_sensor_profiles_demo.py

This script:

• Generates synthetic data for:
  • depth
  • lidar
  • radar
  • thermal
  • IMU (t, gx, gy, gz, ax, ay, az)
• Applies profiles (light, privacy, ghost, chaos)
• Prints summary statistics:
  • trusted stats
  • veiled stats
  • mean and max absolute difference

📦 Coming Soon (v1.0 Release)

• pip package (pip install data-veil)
• Core API expansion
• Integration examples for robotics / cloud pipelines
• Untrusted-access filter templates

📜 License

MIT License

🤝 Contributing

Pull requests welcome from:

• robotics engineers
• cybersecurity analysts
• red-team researchers
• OSINT analysts
• simulation experts

🎉 Thank You

Explore, remix, collaborate, and build on Data Veil.
Designed for creativity, research, and future experimentation.