Drone Swarm Intelligence for Modern Warfare

A drone swarm is a group of autonomous drones that work in coordination and communicate with each other to achieve a common objective. Drone swarms are composed of multiple drone types, ranging from small quadcopters to larger fixed-wing drones. Their strength lies in their ability to function as a collective unit, leveraging swarm intelligence algorithms and advanced software to operate seamlessly and effectively.

Drone swarming technologies are being developed by many countries—particularly major military powers—as part of their efforts to modernize warfare. Although scientists often invoke humanitarian applications such as disaster relief and ecological surveys when discussing this research, it clearly has military potential. As drones become more powerful and affordable, military forces can use them to track and kill adversaries, and potentially even disrupt their weapons systems, communications, and supply chains.

In a drone swarm, the drones communicate with each other via wireless connections or onboard processors, enabling them to adapt their behavior in response to real-time information. For example, drones equipped with cameras and other environmental sensors can identify potential targets, environmental hazards, or defenses and then relay this information to the rest of the swarm. The swarm can then maneuver to avoid the hazard or defense, or weapon-equipped drones can attack the target.

This dynamic behavior can be achieved through a variety of algorithms that control the drones’ navigation, imaging, and decision-making. In particular, our method combines sensor data with image analysis to improve target detection. This is aided by an adaptive sampling strategy that considers both local occlusion density and target view obliqueness, which increases visibility for a given scanning time. We also use a particle swarm optimization (PSO) algorithm to guide the swarm’s behavior, which enables us to optimize performance by minimizing scan time and maximizing the overlapping area covered by the swarm’s camera field of view.

We demonstrate the effectiveness of our approach by analyzing experimental results from a number of scenarios. We found that for a given scanning distance, the optimal PSO solution converges on a stationary target within 14 s, and achieves maximum target view coverage (MTV) of 72% when using a 20-s scan time. Moreover, for a fixed SA, we find that the wider SA and denser sampling of larger swarms lead to better visibility than smaller swarms.

This study demonstrates that drone swarms are feasible for CBRN operations, but additional research is needed to address key challenges. For instance, the Department of Defense should expand its ongoing research into fundamental drone swarm capabilities to include CBRN-relevant uses, such as targeting chemical or biological agents. Ideally, a commander would be able to mix and match drones with different payloads to respond to dynamic battlespace conditions—for example, some drones could perform electronic warfare while others engage hostile forces. In addition, the international community should develop global norms and treaties to limit lethal autonomous drone swarms. To begin, this might mean adding swarm-capable drones to the UN register of conventional arms.