A Multi-Stage Aerospace Engineering Stack: From Structural Foundations to AI-Driven Autonomy.
This repository demonstrates a complete engineering lifecycle. It transitions from traditional aerospace design (ICARUS 1.0) to advanced AI-augmented autonomy (ICARUS 2.0).
- CAD/Geometry: Developed high-fidelity 3.0m wing models in OpenVSP.
- Structural Certification: Performed FEA stress analysis in PrePoMax/CalculiX to certify the airframe under max aero-loads.
- CFD Validation: Conducted aerodynamic simulations in OpenFOAM to establish baseline Lift/Drag polar data.
- Neural Design Optimizer: Integrated a Neural Network (MLP) to programmatically evolve wing geometry, achieving a +12% efficiency gain.
- Shadow Pilot (RL): Developed a MAVLink-AI bridge for real-time attitude control using reinforcement learning principles.
- Digital Twin Health Monitor: Implemented a Random Forest classifier for predictive maintenance and flutter detection.
PrePoMax FEA results showing stress distribution on the HALE airframe under max aero-load.
Neural Network convergence plot showing +12% lift efficiency gain through evolved wing geometry.
| Parameter | Baseline (Initial) | Optimized (Final) | Improvement |
|---|---|---|---|
| Wing Span | 3.00 m | 3.16 m | +5.3% |
| L/D Ratio | 14.2 | 15.9 | +11.9% |
| Max Stress | 9.77e4 Pa | 9.42e4 Pa | -3.6% |
Real-time mission telemetry featuring the Shadow AI Attitude Controller and Structural Health Monitor.
To run the ICARUS optimization and health monitoring stack:
- Python: 3.10+
- Environment: ArduPilot SITL / Mission Planner (optional for flight)
- Tools: OpenVSP (for geometry visualization)
-
Clone the repository:
git clone https://github.com/yogesh031020/Project-ICARUS.git cd Project-ICARUS -
Install core dependencies:
pip install numpy pandas scikit-learn pymavlink matplotlib torch
-
Run the AI Design Optimizer:
python icarus_optimizer.py
-
Launch the Health Monitoring System:
python icarus_health_monitor.py
icarus_brain.py: Core Neural Network logic for aerodynamic prediction.icarus_ai_pilot.py: MAVLink-based attitude control interface.icarus_health_monitor.py: Random Forest classifier for structural failure detection.geometry/: OpenVSP (.vsp3) source models for the airframe.results/: High-resolution engineering data and visuals.
Developed by Yogesh E S - Aeronautical Engineer.