Santa Clara University
A research initiative developing verifiable, safety-critical autonomy for small UAVs operating in human environments.
Project Skyline develops safety-critical autonomy frameworks for aerial delivery systems. Our work integrates control theory, perception, and simulation-based validation to improve UAV reliability in real-world conditions.
We focus on verifiable, failure-resilient autopilot and perception modules for GPS-denied and disturbance-rich environments such as university campuses, industrial sites, and suburban corridors.
| Domain | Tools / Frameworks | Description |
|---|---|---|
| Simulation | MATLAB · Simulink · Gazebo | 6-DoF digital twin and Monte Carlo regression testing |
| Control | PX4 · Control Barrier Functions | Supervisory safety layer for enforcing flight constraints |
| Perception | Stereo Vision · LiDAR · IMU · GPS Fusion | Multi-sensor perception pipeline for robust navigation |
| Analysis | Python · ROS · OpenCV | Data logging, safety metrics, and post-flight validation |
- Autopilot Safety and Certification Readiness – redundant control architectures and verified safety guarantees.
- Obstacle Detection and Sensor Fusion – robust perception under uncertain sensing conditions.
- Simulation-Based Validation – Monte Carlo trials for wind, mass, and sensor fault modeling.
- Operational Risk Modeling – quantitative risk estimation and envelope analysis.
/matlab/autopilot_validation.m— Monte Carlo flight envelope analysis/gazebo/sensor_fault_tests/— randomized LiDAR and IMU fault injection/analysis/safety_metrics.py— probability of constraint violation estimator
- Safety metrics computed over 6,000+ simulated flight scenarios
- Validated control-barrier-function enforcement in real time
- Modular sensor fusion architecture ready for hardware-in-the-loop testing
Burak Kürkçü, Ph.D. — Faculty Advisor
Nikhil Ranjit — [email protected]
Victor Joulin-Batejat — [email protected]