Designing a new wildfire response that uses swarms of drones

This long read is written in collaboration with Mohammad Tavakol Sadrabadi from Coventry University’s Autonomous Vehicle and AI Lab; and Guillermo Rein and Joaquim Peiró both from Imperial College London.

The authors will be talking about this topic at The Emergency Tech Show 2025.

---

Wildfires are increasing in frequency, intensity, and duration across the globe, including in the United Kingdom (UK). Although wildfires in the UK are typically smaller in scale than those in Australia, the United States, or some European countries, their occurrence is rising due to hotter, drier summers and the changing nature of land use.

Fires in moorlands, peatlands, and heathlands can burn for weeks, releasing significant greenhouse gases and threatening communities, critical infrastructure, and biodiversity. With 168 wildfires of at least 30 hectares and a total burned area of 35,000 hectares by late May, 2025 is already a record-breaking year in the UK [1].

As climate change reshapes our fire landscape, emergency services must evolve. One of the most promising tools is the unmanned aerial vehicle (UAV), commonly known as the drone.

In our recent study published in the International Journal of Disaster Risk Reduction [2], we presented a conceptual design for a Wildfire Emergency Response System (WERS) empowered by swarms of drones.

This article sets out how a fleet of self-coordinated drones can support tasks within every phase of the wildfire emergency response: i.e., the pre-fire, active-fire, and post-fire phases. It offers an overview of our work, highlighting how these ideas and cutting-edge technology can be applied to wildfire emergencies in the UK as an integral component of the response system.

Why use drone swarms for UK wildfire emergencies?

While drones are already employed by some UK fire and rescue services for tasks such as surveillance and aerial mapping, most deployments remain isolated and limited in scope. The capabilities of standalone, remotely operated drones are inherently constrained, particularly in dynamic, large-scale, or time-critical scenarios.

We propose a comprehensive system where drones operate as intelligent swarms. They work autonomously and collaboratively in real time to support the full emergency cycle. The aim is for the drone swarm to be an integral component of the system already at the design stage, rather than be added afterwards to handle specific tasks and having to adapt its operations to an existing system that had not been designed to host it.

Thanks to recent advances in autonomy, sensors, swarm robotics, artificial intelligence (AI), and mesh communication networks, such a system is now technically feasible. The use of AI-enabled drones [5] and drone swarms [6] is progressively gaining traction.

A well-designed drone swarm provides fast response times, even for otherwise inaccessible areas, and it can collect environmental data, guide evacuations, provide high-resolution monitoring and situational awareness of fire front and burned area, detect re-ignition risks, support fire propagation forecasts and suppression activities [4] [3], and even support fire investigation—all without placing humans in danger.

This technology is especially relevant to the UK, where many fires burn in hard-to-reach rural terrain, often on conservation land or around scattered housing at the wildland–urban interface (WUI).

Introducing WERS: Wildfire Emergency Response System

The wildfire emergency response is composed of three key operational phases [6]:

• Pre-fire response (preparedness).
• Active-fire response (e.g., suppression, monitoring, predictions, evacuations).
• Post-fire response (e.g., loss evaluations, emergency stabilisation, recovery efforts, seeding)

The diagram below shows a range of tasks or missions which may be carried out or supported by drone swarms within each of these three phases. Mission-specific sensors may be required. The drones are self-coordinated to achieve the swarm’s objectives via autonomous and collaborative decision-making, with human operators overseeing (i.e., assigning tasks, informing, steering, overriding) autonomous operations via intuitive and preferably immersive interfaces.

Having the drone swarms integrated into the WERS means that the same swarm may perform more than one mission, whether simultaneously (as above) or at different instances.

Furthermore, swarm systems are scalable, which means their sizes may change during operations without losing functionality. This makes it possible, for example, for a few drones in a swarm suppressing or tracking a fire front to be lost during operations or temporarily re-allocated to a different mission (e.g. delivery of essential supplies), while the reduced swarm continues to carry out the original mission with only a graceful degradation of performance. Conversely, scalability also allows to seamlessly increase the swarm size during operations as required, gradually adding drones or even merging swarms.

While the same swarm may be used across missions, different drone types may be more convenient for others. For example, heavy-lift drones may be suitable for fire suppression, fixed-wing drones for long-range inspections, while small multi-rotors may be preferred to monitor evacuations.

Pre-fire response: preparedness

In the UK context, drone swarms can be deployed regularly over sensitive landscapes such as the Pennines, the New Forest, or the Cairngorms to monitor fuel accumulation, dryness, and vegetation types. Drones equipped with multispectral sensors can identify changes in fuel conditions before they become a hazard.

Partnerships with local land managers, the National Trust, and conservation agencies can support the routine use of drones to create detailed fuel maps. These maps would improve predictive fire models used by the Met Office and fire and rescue services.

Drone swarms may patrol large areas during periods of elevated risk (e.g., red or amber Fire Severity Index days), warning of unauthorised campfires or barbecues, which are frequent sources of ignition in UK moorland fires.

Active-fire response: enhancing firefighting capacity

The UK’s wildfires often require coordination between several local fire and rescue services, especially in regions like Wales or northern Scotland, where topography and limited access can delay response. When a fire breaks out, the WERS will provide fast response, deploying self-organising drones to:

• Provide thermal and visual imagery of the fire edge
• Identify nearby communities and routes at risk
• Deliver public address messages where conventional communication may be down
• Monitor weather conditions locally (wind shifts and humidity) that affect fire behaviour
• Actively participate in fire suppression activities (including the construction of fire breaks)
• Deliver essential supplies to stranded people and/or firefighters
• Monitor evacuation of pedestrians, vehicles, and assets.

These drones can act as aerial scouts in difficult moorland terrain where fire crews on foot have limited visibility. Instead of waiting for a helicopter or satellite image, a swarm of drones could provide situational awareness, including live video and fire spread forecasting, thus enabling dynamic decision making at the incident command level.

Drone swarms can also act as communication relays in remote areas, self-organising to form ad-hoc survivable networks, supporting both firefighters and landowners on the ground.

Recovery and post-fire monitoring

Drones remain useful even after the fire is suppressed. In the UK, post-fire concerns often focus on peatland degradation, water contamination, and habitat loss. Drone swarms can be deployed to:

• Monitor potential flare-ups in deep-burning peat
• Assess burn severity and identify remaining hotspots
• Assess damage of critical infrastructure
• Track vegetation regrowth and erosion
• Carry out precision aerial seeding
• Provide data for environmental restoration planning and insurance claims.

Drone imagery collected after the Saddleworth Moor fire in 2018 proved invaluable for ecological assessments. We propose formalising this kind of data collection as part of a structured post-fire drone programme, feeding information back into preparedness and land management strategies.

Four key design principles for emergency UAV systems

Designing a usable, effective drone swarm system for wildfire emergency response—whether in the UK or elsewhere—requires more than just good hardware. Our proposed WERS emphasises five key design principles:

• Interoperability. The system must work across different fire services, drone platforms, and data systems. Common standards and open architecture are essential.
• Human–Machine Teaming. The system must augment, not replace, human decision making—even if some tasks are performed autonomously. Intuitive dashboards and control interfaces must empower firefighters and incident commanders.
• Scalability. Whether a two-hectare grassland blaze or a 1,500-hectare moorland fire, the system must scale with the need—deploying the number of drones that the situation requires. This includes changes in swarm size during operations without loss of functionality.
• Robustness. The system must maintain functionality despite disturbances, uncertainties, or variations in operating conditions—e.g., it must be able to handle noise in sensor readings, communication delays, environmental changes, or minor deviations from expected parameters without significant performance degradation.
• Resilience. The system must have the capacity to recover from failures and faults and to adapt to significant disruptions, thus ensuring mission continuity. Fire conditions are harsh. The system must continue to function amid smoke, loss of signal, or individual drone failures.

Legal and ethical considerations in the UK

The use of drones in emergency response raises regulatory and ethical questions, especially in the UK where Civil Aviation Authority (CAA) rules limit autonomous and beyond visual line-of-sight (BVLOS) operations. Any operational WERS deployment must comply with the UK’s aviation safety framework, and collaboration with the CAA will be key.

Public trust is also vital. UK residents have shown concern about drone surveillance in other contexts. For drones to be widely accepted in fire emergencies, transparency and public benefit must be clearly communicated. Data privacy, storage, and sharing policies must be developed and explained.

Call for collaboration

The WERS is not a commercial product, but a design concept backed by technical feasibility and operational logic. Bringing it to life will require collaboration between fire and rescue services, research institutions, regulatory bodies, and drone technology providers.

The UK is well-positioned to lead in this space. We have strong academic expertise, capable fire services, and a landscape where even small-scale UAV deployment could provide substantial benefit. A pilot project to trial coordinated drone swarms in fire-prone areas would be a natural next step.

We would welcome conversations with fire and rescue services on this topic. Please email us: Mauro.S.Innocente@coventry.ac.uk and g.rein@imperial.ac.uk.

Conclusion

Drone swarms will not replace the courage or experience of UK firefighters, but they can serve as powerful allies. Whether supporting fire prevention in peatlands, monitoring fire spread in real time, or assessing post-fire environmental damage, UAVs are a scalable and intelligent tool for modern emergency response.

As we face a future with more frequent and intense wildfires, including in parts of the UK once thought safe, embracing such innovation is not optional—it is essential.

References

[1]   “EFFIS Annual Statistics for United Kingdom,” Copernicus. Europe’s eyes on Earth, [Online]. Available: https://forest-fire.emergency.copernicus.eu/apps/effis.statistics/estimates/GBR. [Accessed 02 06 2025].

[2]   M. Tavakol Sadrabadi, J. Peiró, M. S. Innocente and G. Rein, “Conceptual design of a wildfire emergency response system empowered by swarms of unmanned aerial vehicles,” International Journal of Disaster Risk Reduction, vol. 124, no. 105493, 2025.
https://doi.org/10.1016/j.ijdrr.2025.105493

[3]   M. Innocente and P. Grasso, “Self-organising swarms of firefighting drones: Harnessing the power of collective intelligence in decentralised multi-robot systems,” Journal of Computational Science, vol. 34, pp. 80-101, 2019.
https://doi.org/10.1016/j.jocs.2019.04.009

[4]   M. Tavakol Sadrabadi and M. S. Innocente, “Enhancing wildfire propagation model predictions using aerial swarm-based real-time wind measurements: A conceptual framework,” Appled Mathematical Modelling, vol. 130, p. 615–634, 2024.
https://doi.org/10.1016/j.apm.2024.03.012

[5]   S. Boroujeni, A. Razi, S. Khoshdel, F. Afghah, J. Coen, L. O’Neill, P. Fule, A. Watts, N.-M. Kokolakis and K. Vamvoudakis, “A comprehensive survey of research towards AI-enabled unmanned aerial systems in pre-, active-, and post-wildfire management,” Information Fusion, vol. 108, no. 102369, 2024.
https://doi.org/10.1016/j.inffus.2024.102369

[6] H. Karvonen, E. Honkavaara, J. Röning, V. Kramar and J. Sassi, “Using a semi-autonomous drone swarm to support wildfire management – A concept of operations development study,” in Engineering Psychology and Cognitive Ergonomics, Springer Nature, 2023, p. 235–253.
http://doi.org/10.1007/978-3-031-35389-5_17

---

Read how Kent Fire and Rescue Service is preparing to deal with the increasing number of wildfires – New wildfire kit for Kent as climate risk continues to grow.