On November 3, 2026, the Computing Community Consortium (CCC) hosted its latest Community Chat, Computing on the Fly: Navigating a Vision for the Future of Drone Computing, based on the recent workshop report of the same name. CCC Executive Committee member and lead Computing on the Fly author Kevin Butler (University of Florida) opened the webinar, fellow lead author Christopher Stewart (The Ohio State University) moderated, and Sabine Brunswicker (Purdue University) and Dirk Giles (USDA Forest Service) gave lightning talks.

Watch the Recording Here

A Roadmap to 2035

Christopher Stewart described the December 2025 Computing on the Fly workshop, where 47 experts from academia, industry, and government outlined a vision for non-military drone applications in 2035, from urban and indoor delivery to medical supply transport and wildfire detection. In the wildfire scenario, networks of drones would monitor hundreds of thousands of acres, assess emerging risks, and help direct firefighters to where they are needed. The impact of such a fleet would be transformative

To get there, the report identifies 12 technical challenges. Stewart highlighted four:

• Closing the simulation-to-reality gap, so software tested on simulated fleets behaves the same way in the field
• Integration testbeds that support reproducible results, graduated complexity, and realistic failure injection
• Scalable identification infrastructure to track what is in the airspace and whether it is behaving as expected
• Real-time autonomy frameworks that adapt fleets at runtime, for example as a mission grows from 5 drones to 50

He also pointed attendees to a forthcoming vision paper on Computing on the Fky in IEEE Internet Computing and an open call for the magazine’s special issue on drone computing.

From Simulation to Digital Twins

Sabine Brunswicker, founder of Purdue’s AIrTonomy program, opened with footage from a recent flight of a heavy-lift drone with a 32-foot wingspan. She argued that the field needs to move past high-fidelity simulation toward digital twins: models of specific physical systems, updated with real-time data, that can run “what-if” analyses during flight.

Her examples included:

• An engine digital twin that detects anomalies from sparse sensor data, now being tested with an industry partner
• Language models constrained to the controlled grammar of air traffic control, which is still largely voice-based, to reduce operator workload and increase safety
• A browser-based simulation environment, used in a recent competition, that lets researchers test algorithms

She closed by inviting labs, testbeds, and test sites to share data and models so the community can build these environments together.

Scaling Drones at the U.S. Forest Service

Dirk Giles, National UAS Branch Chief at the USDA Forest Service, described how the agency has scaled its drone program on the same standards-based approach it applies to piloted aircraft, with publicly available standards for training, operations, airworthiness, and pilot qualification.

On wildfires, drones give crews a thermal view for finding spot fires beyond containment lines, and they are taking on aerial ignition, which the agency has conducted for more than 50 years from low-flying helicopters at real risk to crews. The agency also uses drones for timber, hydrology, reforestation, and infrastructure inspection.

Data management has been just as important as the aircraft. The agency processes and stores drone imagery in line with federal records requirements. That data now feeds AI tools for change detection in bridge inspections and for counting nursery seedlings, with 100% accuracy for the youngest plants, replacing manual inventories.

Looking ahead, his priorities are safe deconfliction with piloted aircraft and the responsible use of taxpayer funds. He also anticipates applications for aircraft with greater endurance, payload, and range, noting that current battery technology cannot yet replace a helicopter.

Q&A: Security, Digital Twins, and the Human in the Loop

Report co-authors Ufuk Topcu (University of Texas), Deborah Silver (Rutgers University), Weisong Shi (University of Delaware), and Nils Aschenbruck (Osnabrück University) joined Butler, Stewart, and the guest speakers for a live Q&A. Some of the topics they discussed included:

• Security. Giles pointed to new federal requirements for drone security and supply chains. Brunswicker noted that new communication systems such as Remote ID make assured autonomy a priority.
• Digital twins in the field. Shi described how real-time digital twins can help locate drones that fail mid-mission in remote areas such as forests. Brunswicker added that true digital twins depend on shared real-world data, which requires collaboration across academia, industry, and policy.
• Broadening participation. Aschenbruck observed that drone experiments are increasingly difficult for mid-size universities to run, and that shared digital twins could open this research to more institutions. He and Silver noted that IEEE and many of its societies, including those focused on edge computing and standards, are expanding into this space.
• Incorporating AI. Asked how close AI-driven drones are to everyday practice, Giles said a human will remain in the loop for the rest of his career, and that the greater need is tools that support pilots in busy airspace. Brunswicker described the goal as augmenting operators without overloading them. Stewart added that in a fast-growing field, an ambitious vision should not be mistaken for an unlikely one.

Learn More and Get Involved

We encourage the computing community to read the full Computing on the Fly report and to look for the authors’ forthcoming vision paper in IEEE Internet Computing. Watch the full recording of the Community Chat on YouTube here.

Tune in to the CCC LinkedIn Showcase Page for updates and more Community Chats like this. Stay connected with CCC for the latest insights, publications, and opportunities to engage by subscribing here.

The Computing on the Fly: Navigating a Vision for the Future of Drone Computing workshop was generously supported by the IEEE Computer Society.

This material is based upon work supported by the U.S. National Science Foundation (NSF) under Award Nos. 2300842 and 2619366. These awards support the Computing Community Consortium (CCC), a programmatic committee of the Computing Research Association (CRA). Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.