This project aims to secure intra-vehicle communications using a dual-layer model combining image-based and data-based approaches. By leveraging deep learning and structured data analysis, the system can effectively detect and prevent cyber threats, ensuring robust vehicle network security.
demo.1.mp4
- Pattern Recognition in Visual Data: Utilizes deep learning to analyze visual representations of network activity, such as CAN bus signal patterns or heatmaps, to detect unusual shapes or spikes that may indicate cyber threats.
- Enhanced Detection of Anomalies: Identifies deviations in normal waveform patterns to spot attacks that alter typical signal flow, such as spoofing or injection attacks.
- Structured Data Analysis: Processes network data (e.g., frequency, timing, and sequence of sensor readings and messages) to detect irregularities that could signal attacks like Man-in-the-Middle (MITM) or Denial of Service (DoS).
- Adaptable to Data Types: Can analyze various structured data points from different Electronic Control Units (ECUs), allowing for a versatile approach across multiple vehicle components and communication protocols.
- Cross-Model Validation: Combines outputs from both models to strengthen accuracy and minimize false positives, ensuring comprehensive detection across image and structured data.
- Robust Against Complex Attacks: Capable of detecting both visual and non-visual anomalies, covering a wide range of potential intra-vehicle threats.
- Fuzzy Attack: A cybersecurity testing method where random, malformed, or unexpected data is fed to software, systems, or devices to identify vulnerabilities and weaknesses.
- Denial of Service (DoS) Attacks: Overloads the vehicle network with false data to disrupt essential functions like braking or steering control.
- Spoofing Attacks (RPM and Gear): Injects false signals or messages into the network, impersonating legitimate components to mislead the vehicle system.
- Subtle, Low-Signal Attacks: Sophisticated attackers often create small, unnoticeable anomalies, which are hard to distinguish from normal fluctuations in high-speed data flows.
- High Data Volume: Vehicle networks generate and exchange massive amounts of data in real-time, making it difficult to identify attacks without causing system delays.
- Time-Sensitive Processing: Attack detection must happen instantly to be effective, requiring highly optimized algorithms that balance security with processing speed.
| MODEL NAME | ACCURACY |
|---|---|
| CNN | 0.9762 |
| XCEPTION | 1.00 |
| VGG16 | 0.9989 |
| VGG19 | 1.00 |
| RESNET | 0.98 |
| INCEPTION | 0.9996 |
| INCEPTION RESNET | 0.9999 |
| MODEL NAME | ACCURACY |
|---|---|
| Logistic Regression | 0.9957 |
| KNN | 0.9999 |
| Random Forest | 1.00 |
- Programming Language: Python
- Deep Learning Frameworks: TensorFlow, PyTorch
- Visualization Tools: Matplotlib, Seaborn
- Dataset: Car Hacking Dataset
- Frontend: HTML, CSS, JavaScript
- Backend: Flask
- Clone the repository:
git clone https://github.com/Rijul1607/V-Secure.git
- Navigate to the project directory:
cd V-Secure - Install the required dependencies:
pip install -r requirements.txt
- Data Preprocessing: Use the provided scripts to preprocess both image and structured data.
- Model Training: Train the image-based and data-based models using the preprocessed data.
- Real-Time Monitoring: Deploy the dual-layer model for real-time intra-vehicle communication monitoring.
Access the DataSet here: Database Repository
Contributions are welcome! Please feel free to submit a Pull Request.
This project is licensed under the MIT License. See the LICENSE file for details.
For any questions or inquiries, please contact Rijul.