Autonomous drones could deliver packages, inspect bridges and skyscrapers, monitor emergencies like wildfires and, eventually, ferry people. But the airspace still lacks the testing and coordination infrastructure needed to prove drones can operate safely around people, buildings, aircraft and each other.

Mehrnaz Sabet, a doctoral student in the field of information science in the Cornell Ann S. Bowers College of Computing and Information Science, is working to create that infrastructure.

Credit: Louis DiPietro/Cornell Bowers

Sabet is the principal investigator behind Project Orion, an autonomous airspace coordination system built around a real-time simulation and validation technology she developed at Cornell. The platform combines simulation, field testing, vehicle-to-vehicle communication and hardware-in-the-loop validation, a technique integrating real drone hardware with simulated airspace environments. The project seeks to test autonomous drone operations under realistic conditions, train their onboard artificial intelligence models and prepare them for safe, autonomous flight.

“Airspace as we know it is going to change,” said Sabet, who has led a team of Cornell graduate and undergraduate students on the project for the last two years. “Many types of drones are going to be integrated into the airspace. We need new infrastructure that, like the drones themselves, is highly automated and autonomous.”

Project Orion has received support from NASA and the National Science Foundation, and Sabet has engaged with federal agencies and industry stakeholders including the Federal Aviation Administration, National Institute of Standards and Technology (NIST), Qualcomm Technologies, Inc. and commercial drone operators. Last month, Sabet published a technical report for government and industry, drawing on Orion’s simulations and tests. 

The project allows physical drones flying in an open test field to interact in real time with a simulated city-sized airspace, including virtual aircraft, buildings, obstacles, emergency scenarios and degraded communication conditions. This mixed-reality setup allows the team to test scenarios that would be difficult, expensive or unsafe to reproduce directly in public airspace.

“Our goal is to close the gap between simulation and deployment,” Sabet said. “Offline simulation is useful, but it does not fully capture what happens when real hardware, onboard autonomy, communication links and live operations interact in real time.”

This year, NIST licensed the underlying simulation technology for public-safety drone research. Additionally, the Commercial Unmanned Aerial Vehicle Expo’s 2026 Innovation Spotlight, an industry-facing showcase, highlighted Orion and its real-time validation infrastructure for drone operations. NASA, too, featured Project Orion in a recent writeup.

“Mehrnaz sees a future where drones coordinate with other drones and human operators to perform many tasks central to society, including search and rescue, security and delivery, and she’s brought together an incredible variety of resources to make this vision happen,” said Susan Fussell, professor of information science and communications and Sabet’s co-adviser with Sanjiban Choudhury, assistant professor of computer science. “Her hard work has transformed how we think about and support drones across all areas of society.”

In collaboration with Qualcomm, the Orion team is testing an early prototype of airborne vehicle-to-vehicle communication for drones, exploring how autonomous aircraft can exchange information directly to support future coordinated airspace operations.

“With Project Orion, Mehrnaz and her team fostered an idea, built something of real value and advocated for it,” said John Cavolowsky, director of NASA’s Transformative Aeronautics Concepts Program, which seeded Sabet and Project Orion with $80,000 in 2024 through its University Student Research Challenge. “The tools, simulations and high-end computing assets she’s advancing will prove critical in the development of future air traffic management systems.” 

Researchers have also used Orion to reconstruct real-world drone incidents. After a delivery drone collided with a construction crane in Arizona in 2025, the team recreated the site in Orion’s validation environment to study how similar hazards could be detected, communicated and tested safely before deployment.

“Real-world incidents show us where the hard cases are,” Sabet said. “Orion lets us turn those incidents into repeatable scenarios, so we can evaluate how autonomy and communication systems should respond before similar operations are scaled.”

They’re extending the same infrastructure to public-safety scenarios, including wildfire monitoring and emergency response missions where multiple drones may need to operate near crewed aircraft, changing hazards and limited communication coverage.

“We are building a proving ground for the next generation of airspace autonomy,” Sabet said. “Before drones become part of everyday life, we need evidence that the autonomy, communication and safety systems work together under realistic stress. That is what Orion is designed to provide.”

Louis DiPietro is a writer for the Cornell Ann S. Bowers College of Computing and Information Science.