Controlling the airspace for small package delivery by uncrewed aircraft systems, or UAS, will require careful planning, informed by modeling and analysis.

On a recent day in Howard County, Maryland, the skies were abuzz with package and meal deliveries. More than 15,000 small UAS flights took off from two warehouses and three fast-food outlets throughout the day. But problems cropped up where the airspace was too crowded with drones. French fries got cold and packages were delayed. The average delivery arrived 52 minutes late. Nearly eight percent of all deliveries were running so late that they were canceled.

Fortunately, no one in Howard County went hungry that day, because the deliveries all took place within a simulation by CNA’s UAS Cooperative Airspace Traffic Simulation tool, or UCATS™. While drones aren’t routinely delivering packages to our doorsteps yet, the Federal Aviation Administration and all government and industry stakeholders must prepare for this reality. Modeling and simulation are essential tools to help smooth the path to this future.

Who will manage the drone delivery airspace?

The entrance of new uncrewed vehicles into the airspace poses a unique challenge for the FAA and industry, partly because the volume of UAS operations is expected to far exceed that of traditional aircraft. Since 2016, over 900,000 UAS have been registered with the FAA—nearly three times the number of traditional aircraft that have been registered with the FAA since its founding in 1958. One study has forecasted that between 8 and 86 million UAS package deliveries will be flying each day in the US within 20 years. As opposed to traditional air traffic management, which is performed centrally by the FAA, UAS traffic management will fall mostly on the UAS operators and service providers themselves. According to the FAA, drone operators, their supporting services, and the FAA will work cooperatively to ensure the safety of the airspace, but drone operators will be responsible for managing the safety of their flights without direct FAA control.

Without a centralized regulator to ensure equitable use of the airspace among operators, the UAS ecosystem must be managed in a way that allows various types of competing operators to access the airspace in a fair manner. Innovative tools are needed to identify key factors and behaviors that impact the fair use of the airspace.

Modeling for fair and timely UAS deliveries

CNA’s UCATS tool models the flight planning process for drones. Our model assumes that operators file flight plans specifying the departure time and trajectory of each delivery. Using an agent-based modeling approach, UCATS tracks individual operations through the process of proposing a flight plan, identifying whether conflicts in time and space occur, and deconflicting as needed before submitting a final plan. The model can then record whether each flight was (1) accepted as planned, (2) delayed to avoid conflicts, or (3) canceled because the system could not support the desired flight plan in time.

We designed a series of scenarios to evaluate a full day of flights in Howard County, running each scenario a thousand times to explore the range of possible outcomes. In our scenarios, we simulated operations from two package delivery warehouses at different demand levels for deliveries and varying airspace density restrictions. We also ran a scenario with three additional fast-food delivery operators servicing the residents in the county.

With regards to fairness, we found that using a first-filed-first-served approach would underprioritize flights that cannot be submitted far in advance and flights that departed later in the day. These flights would have a greater chance of incurring long delays or being canceled. In particular, meal orders, which are rarely made far in advance—and are time-sensitive—could not compete fairly with package orders in a first-filed-first-served approach. When we took package delivery out of the scenario, almost no meal orders were canceled. This suggests that a segregated airspace may help improve equity among operators that have short file-ahead times, such as fast-food delivery services.

Modeling and analysis for the future airspace

With our initial case studies, we have only started to investigate the many questions that arise around widespread integration of UAS operations. UCATS gives decision-makers and analysts a tool to quantitatively evaluate what-if scenarios to provide insights into the behavior of the future airspace.

As the drone industry and regulators continue to plan for the future, innovative tools and approaches are needed to ensure the future airspace is safe and fair. UCATS is one decision support tool that can be used to provide data-driven insights, and CNA looks forward to exploring other timely questions around UAS integration using UCATS. We also hope to expand this type of simulation approach to other use cases, such as air taxis, and to entirely new systems that can benefit from agent-based analysis.

To read additional details and results from our study, download the report, Agent-Based Modeling of Uncrewed Aircraft System Flight Planning for Airspace Fairness. To hear our analysts discussing the approaching integration of small drones or uncrewed aircraft systems (UAS) into the US airspace, listen to our recent CNA Talks podcast. To read about another agent-based modeling tool that CNA developed to model virus spread in correctional facilities, visit the SAFER-C™ information page. 

Rebekah Yang is a systems engineer in CNA’s Center for Data Management Analytics supporting the future integration of emerging technologies, such as UAS and artificial intelligence, for the FAA.

Adam Monsalve is a systems engineer in CNA’s Center for Enterprise Systems Modernization specializing in the intersection of uncrewed technologies and cybersecurity for multiple federal government clients.

Mark Lesko is a research scientist in CNA’s Center for Data Management Analytics specializing in issues and analysis related to traditional air traffic management.