Significance
Although robust corrections for GPS signals exist at ground-level and high-altitude ranges, there remains a critical gap at low to mid altitudes—especially in urban canyon environments where signal obstruction, reflection, and distortion are common. Addressing this gap is crucial for next-generation aerospace applications, drone operations, and autonomous navigation in congested cities. By pioneering new correction algorithms and hardware tailored for these challenging settings, our research fills this unmet need, enabling safer and more reliable localization for diverse commercial and research endeavors. This project is currently being commercially developed and is undergoing deployment testing.
Fusion Flight
Overcoming Challenges and Research Milestones
1
Implementing Complex GPS Corrections
IRNN-DNN (Iterative Recurrent Neural Network–Deep Neural Network) employs iterative refinement within a deep learning framework to correct GPS signals dynamically. It integrates domain knowledge about multi-path effects and leverages satellite-obstacle geometry to provide robust GPS error corrections.
2
Versatile Platforms for Field Testing
PATHFINDER and SCOUT hardware platforms (Plane, drone and near-earth space vehicle configurations) are designed to accommodate different payloads and sensing packages. They enable comprehensive data collection under diverse conditions—ranging from urban canyons to open fields—facilitating algorithm validation and iterative improvements in real-world scenarios
3
Hardware Integration, Cost, and Reliability
Custom-designed boards and microcontrollers handle sensor fusion and real-time processing, ensuring cost-efficiency and streamlined integration