A sophisticated real-time 3D navigation system designed for autonomous drones operating in GPS-denied environments. This project was funded through the Advanced Technology Research Center (ATRC).
Technical Architecture
- Custom Extended Kalman Filter (EKF) for sensor fusion combining IMU and camera data
- Feature-based visual tracking using ORB descriptors with RANSAC outlier rejection
- Tightly-coupled sensor fusion achieving sub-centimeter accuracy
- Real-time loop closure detection for drift correction
- Optimized C++ implementation running at 30+ FPS on embedded hardware
Key Achievements
- 5% reduction in computational complexity while maintaining precision
- Successfully deployed on custom drone platform
- Published research at IEEE CSCI 2023
Funding
Advanced Technology Research Center (ATRC) under the Institute of Technology