- A new peer-reviewed study from Kuwait is exploring how drone swarms could make firefighting safer
- It sets out a six-drone framework designed specifically for firefighting operations
- The authors focus on two key risks that firefighting drones face
A new peer-reviewed study from Kuwait is exploring how drone swarms could one day make firefighting safer and more reliable, even in harsh and hostile conditions where both heat and cyber threats are in play.
The paper, published in the January 2026 edition of the Journal of Computer and Communications, sets out a six-drone framework designed specifically for firefighting operations. Its goal is to keep drone missions running even when individual aircraft begin to degrade in extreme environments, and to reduce the risk of cyberattacks targeting the wireless links that hold drone swarms together.
Authored by Ahad Alotaibi and Abdullah Alrasheedi from the Department of Advanced Technology at the Canadian College of Kuwait, the research reflects growing interest across the Middle East in using autonomous systems to support emergency response while minimising risk to human crews.
At the heart of the proposal, five drones are assigned operational roles, including thermal monitoring, environmental sensing, close-range visual inspection, payload support, and communications relay.
Alongside them is a sixth aircraft, known as the Inspector or Commander drone, which plays a supervisory role. Hovering in a vantage position, it keeps an eye on the rest of the swarm, captures inspection imagery and acts as the main coordination link back to the ground control station.
The authors focus on two key risks that firefighting drones face.
The first is physical damage caused by heat, smoke, turbulence and debris. The second is the growing cyber threat to wireless drone communications, particularly the risk of attackers intercepting or manipulating coordination traffic during a mission.
To tackle physical degradation, the framework introduces an AI-driven inspection process carried out while the drones are still in flight.
The Inspector drone periodically captures high-resolution images of its neighbouring drones and sends them through the swarm network to a mobile application called Drone Inspector. From there, the images are uploaded to a cloud-based AI model using Amazon Rekognition Custom Labels, which analyses them for signs of damage.
The system is trained to spot issues such as exposed wiring, damaged or misaligned landing gear and deformation of key components. Inspection frequency changes depending on risk. Under normal conditions, checks take place every two minutes, but when onboard gas sensors suggest the swarm is close to an active fire zone, inspections increase to every 30 seconds.
If the AI detects a problem with a confidence level above a defined threshold, set at 80 per cent in the study, an alert is sent straight to the ground control station.
Alongside this physical monitoring, the paper proposes a cyber-resilient communications design aimed at making attacks harder to carry out. Rather than running all drones on a single flat network, the swarm is divided into role-based subnets, with strict rules governing how data flows between them.
This segmentation is combined with a dynamic routing approach that constantly adjusts communication paths as conditions change, reducing the chances of attackers finding a stable point from which to intercept traffic.
To test the concept, the researchers carried out a series of evaluations covering deployment feasibility, AI inspection performance and cybersecurity resilience. A six-drone setup matching the proposed architecture was used, with commercially available platforms such as DJI Matrice models and an Autel EVO Max taking on different roles.
For safety reasons, live fire was not involved, but the team used formation flying, manoeuvres and variable wind conditions to recreate the kinds of stresses drones might face during real emergency responses.
The findings suggest that combining in-flight physical inspection with cyber-aware network design could significantly improve the reliability of drone swarms in high-risk environments.
