Swarms of autonomous robots, capable of coordinating and executing complex tasks, are a futuristic vision that is quickly becoming a reality. However, even as these systems grow more sophisticated, they still rely on human oversight. This raises an important question: How many robots can one person manage before becoming overwhelmed?

The Defense Advanced Research Projects Agency’s (DARPA) was worried they could end up in a situation with too many robots and too few people.

So they wanted to see whether a human could effectively supervise a heterogeneous swarm of more than 100 robots, even in challenging real-world scenarios. While brief moments of overload occurred, the human operator successfully managed the tasks overall.

Playing real-time strategy games in real life

The study was led by Julie A. Adams, the associate director of research at Oregon State University’s Collaborative Robotics and Intelligent Systems Institute. It was conducted under DARPA’s OFFensive Swarm-Enabled Tactics (OFFSET) program, which aims to explore the capabilities of robot swarms in complex, tactical scenarios. It’s a bit like controlling units in a real-time strategy computer game, except in real life.

Two human controllers took turns engaging in a series of military mock missions. They were not beginners — the operators were well-trained and had previous experience. They alternated shifts ranging from 1.5 to 3 hours per day. The largest mission involved a whopping 110 drones, 30 ground vehicles, and up to 50 additional vehicles.

The missions required human controllers to assign tasks, navigate a virtual reconstruction of the environment, and manage the swarm’s response to thousands of hazards. These hazards, represented by simplified QR codes, added layers of complexity—some required the coordination of multiple robots, while others moved dynamically throughout the environment.

The researchers didn’t just monitor how well the operators were completing the tasks. They also monitored their heart-rate variability, posture, and speech rate. Researchers developed an algorithm that analyzed all this data and identified when the operator was overworked and unable to satisfy all the demands of the operation.

It’s not the size of the swarm

Previous theories suggested that increasing the number of robots would proportionally raise the human operator’s workload. However, the DARPA study revealed that other factors, such as system design, mission complexity, and task requirements, played a more significant role. Overload moments were often triggered by the need to generate new tactics or reassess task assignments, rather than the sheer volume of robots.

“While the swarm commander’s workload estimates did cross the overload threshold frequently, the swarm commander was able to successfully complete the missions, often under challenging operational conditions. The presented results demonstrate that a single human can deploy a swarm of 100 heterogeneous robots to conduct real-world missions.

Adams and her team are now trying to investigate exactly what factors demand on the operator. While DARPA is interested in military applications, one of the most promising applications for this type of research is actually for disaster relief. Humans could operate robot swarms in hazardous environments where human lives could be at risk. For instance, they could coordinate drones, inspection robots, and other vehicles in wildfires, identifying hotspots and guiding resources in areas that need them the most.

For now, the key lies in designing systems that balance human input with autonomous capabilities. With further research, we could see swarms operating in ways that were once the realm of science fiction, all under the steady guidance of a single human controller.

The study “Can a Single Human Supervise a Swarm of 100 Heterogeneous Robots?” has been published in IEEE Explore.