This repository implements a complete hybrid AI system combining classical machine learning and quantum-inspired optimization to detect gas pipeline leakages and coordinate drone response paths. The system leverages synthetic acoustic signals based on the GPLA-12 dataset and includes visualization, analytics, and animation.
J. Li and L. Yao, “Gpla-12: An acoustic signal dataset of gas pipeline leakage,” 2021.
Gas leaks in pipelines can be hazardous to human health, safety, and the environment. Traditional manual inspection is time-consuming and lacks scalability. This hybrid system enables:
- Scalable Automation: Detect and respond to leaks across large areas using autonomous drones.
- Early Warning Systems: Rapid classification of high-pressure acoustic anomalies.
- AI-Driven Efficiency: Quantum-inspired TSP and Voronoi path planning improve coverage and reduce drone travel cost.
- Explainable Intelligence: Classical feature extraction ensures transparency and model interpretability.
| Domain | Application |
|---|---|
| Smart Infrastructure | Monitor gas pipelines, sewer systems, and pressurized conduits |
| Environmental Safety | Detect methane or hazardous gas emissions in remote or urban environments |
| Disaster Response | Coordinate UAVs for leak inspection after earthquakes or industrial failure |
| Precision Agriculture | Detect toxic gas or irrigation system leaks |
| Research & Simulation | Benchmark quantum-inspired optimization in AI safety systems |
Best Test Accuracy: 93.43% (Classical Neural Network) Hybrid Architecture: 60% Classical + 40% Quantum Sample Count: 684 synthetic acoustic samples across 12 classes
- End-to-end pipeline from data simulation to drone path animation
- Feature-rich classical models: Random Forest, SVM, Neural Network
- Quantum-inspired models for classification and drone optimization
- Quantum Voronoi-based drone region assignment
- Animated visualization of multi-drone mission paths
This project is designed to run in Google Colab or locally with Python 3.8+.
git clone https://github.com/yourusername/hybrid-gas-leakage-detection.git
cd hybrid-gas-leakage-detection
pip install -r requirements.txtpython hybrid_gas_leakage_system.pySteps executed:
- Mount Google Drive (if in Colab)
- Download or generate the GPLA-12 synthetic dataset
- Extract classical and quantum features
- Train models and evaluate performance
- Generate drone paths using quantum Voronoi + quantum TSP
- Create animations and visual comparisons
| Component | Type | Purpose |
|---|---|---|
| Feature Extraction | Classical | Time, frequency, wavelet, and statistical feature engineering |
| Classification | Classical | Random Forest, SVM, Neural Network |
| Classification | Quantum | Quantum-Inspired Classifier using angle encoding + circuits |
| Drone Optimization | Quantum | Quantum Voronoi site placement and TSP path planning |
The system produces a step-by-step animation of the drone mission using Matplotlib’s animation module. The animation illustrates:
- Drone deployment to Voronoi-partitioned zones
- Individual waypoints and paths for each drone
- Leak intensity represented as background scatter
- Total mission progression over time
The animation is saved to Google Drive (if in Colab) as:
/content/drive/MyDrive/datasets/drone_mission_animation.gif
This visual output is essential for debugging drone path logic, analyzing quantum optimization convergence, and presenting mission dynamics clearly in academic or industrial reports.
| Model | Type | Test Accuracy |
|---|---|---|
| ClassicalNN | Classical | 93.43% |
| SVM | Classical | 89.05% |
| Random Forest | Classical | 87.59% |
| QuantumClassifier | Quantum | 11.68% |
| HybridEnsemble | Hybrid | 91.97% |
- 684 synthetic acoustic samples
- 1024-length signals simulating real gas pipeline leakage
- 12 classes derived from combinations of pressure, noise, and microphone
Each signal includes:
- Modulated base frequencies
- Harmonics
- Noise perturbations
- Microphone-specific adjustments
- Model training curves (accuracy & loss)
- Bar charts of model comparison
- Quantum optimization convergence plots
- Voronoi region mapping and drone waypoints
- Real-time drone animation
This project is intended for academic and research purposes only. It is not certified or approved for real-world deployment in safety-critical infrastructure or commercial drone navigation systems. Use at your own risk and ensure compliance with local regulations and safety standards when adapting this code for testing or experimentation.
MIT License