Growing up alongside deadly fires inspired me to study them – and fight flames with swarms of drones
Growing up in Greece, wildfires were a constant presence each summer. In 2007, I remember watching TV footage of fires ravaging the Peloponnese peninsula and island of Evia, destroying forests and homes, taking lives. The sight of helicopters and firefighting aircraft crossing the smoky skies was both terrifying and awe-inspiring.
Then when I was 17, flames crept dangerously close to my home in Kavala, northern Greece. I recall standing outside with water-hose in hand, scanning the horizon and hoping our nearby treeless street would stop the fire's advance. Thankfully, firefighting aircraft reached the area just in time - but the feeling of vulnerability at seeing how easily entire landscapes could be consumed stayed with me.
Those experiences shaped my curiosity about how people could better respond to such disasters. Wildfires are becoming more intense, frequent and harder to manage worldwide as fire seasons become longer, affecting communities from California to Australia.
According to the UN environment programme, longer droughts, heatwaves, and erratic winds are pushing ecosystems past their natural limits, endangering both human lives and biodiversity. Nasa reports that extreme wildfire activity has more than doubled worldwide over the past two decades.
In 2018, Greece suffered the deadliest wildfires in its modern history when fires in the southern seaside town of Mati and in the general Attica region claimed over 100 lives. The devastation renewed my determination to find better ways to combat fires.
The following year, while doing a master's degree in robotics at the University of Bristol, I joined a hackathon event, organised by the UK government's Defence Science and Technology Laboratory and the Lancashire Fire and Rescue Service, about using "swarm intelligence", AI and drones to improve wildfire detection and response.
Swarm intelligence describes the exchange of information by decentralised, self-organised systems in order to solve complex problems. It is inspired by such collective behaviour in nature, for example by flocks of birds or swarms of insects. The competition sparked my interest to investigate how these tools could be used in such potentially catastrophic events.
After the hackathon, my supervisor Sabine Hauert, a professor of swarm engineering, and I were approached by Windracers, a UK company specialising in heavy-lift drones capable of carrying hundreds of kilograms of payloads including water to remote areas.
Transforming my childhood wildfire experiences into tangible technology through a PhD project was irresistible. The challenge was how to develop these drones into a swarm that could be used for quicker and more effective detection of, and response to, potentially catastrophic wildfires.
XPrize challenge
Today, I lead the Aura team (short for Autonomous Ubiquitous Response with Aware Robots), one of 15 semi-finalists in the wildfire section of XPrize - the series of competitions seeking technological solutions for the world's "most urgent and complex challenges". We were also chosen to be one of the Prototypes for Humanity exhibiting in Dubai in 2025.
Aura comprises experts from the universities of Bristol and Sheffield plus members of Lancashire Fire and Rescue. The challenge set by XPrize was simple to describe, but technologically demanding: monitor 1,000 square kilometres of land for a full day and, upon detecting a fire, extinguish it within ten minutes.
Aura's technology stems in part from my PhD research during the pandemic lockdowns. Unable to work in the lab, I reached out to firefighters, foresters and emergency professionals worldwide for insights. Through interviews and focus groups including extensive collaboration with the Lancashire Fire and Rescue Service, who frequently use drones when responding to wildfires, we shaped the swarm system based on real operational needs.
Our approach uses commercially available drones, such as quadcopters, equipped with custom software that transforms them into a coordinated swarm. Like a flock of birds, they operate without a central leader, relying on interactions with one another about their location and other information to continuously adapt to their environment.
This allows a single operator to control multiple drones simultaneously, because the drones perform some tasks safely without any need for human intervention. This is an essential capability for large-scale, rapid responses.
The firefighters guided us on what truly matters in the field: reliability, usability and speed. They emphasised the human challenges of wildfire response: long shifts under extreme heat, difficult terrain, with a constant risk to their as well as other people's lives.
Eradication is not always the answer
By offering firefighters an aerial support team that can scout, map and even deploy extinguishing material autonomously, Aura aims to extend their reach and safety rather than replace their expertise.
The fire practitioners we work with, in the UK and other countries such as Greece and Canada, often remind us their goal is not to eradicate every wildfire. Fire is a natural and necessary element of many ecosystems, so the challenge lies in managing it, preventing small fires from becoming catastrophic ones while allowing controlled burns that sustain biodiversity.
By reducing the amount of vegetation, controlled burns can reduce the intensity of future wildfires. These are practices that people have been using throughout human history, including Indigenous people in North America and Australia.
Related content: Did you know that goats can reduce wildfire risk?
🔥 Did you know that goats can help prevent wildfires?
Our swarming drones system supports that balance by acting as an intelligent tool to help firefighters and land managers make faster, more informed decisions. Our vision is to see drones not only fighting fires but also assisting in disaster logistics: delivering supplies, monitoring hotspots, and supporting crews in the field.
Despite the rapid pace of innovation, however, drone regulation still lags behind technology. In most countries, operating drones "beyond visual line of sight" (BVLOS) requires special authorisation. Dropping payloads of even small amounts of water on a wildfire also involves lengthy safety assessments. These restrictions make testing swarm systems such as Aura challenging.
But progress is on the horizon. Regulators are beginning to approve limited BVLOS operations for certified operators in the US, Australia, Canada and the UK. But a more flexible, data-driven approval process, one that builds cumulative safety cases from successful missions, could unlock greater potential for autonomous systems like ours.
We still have a long way to go to make these technologies a reality, but the ambition that drives me is the one that began when those flames threatened my childhood home. To protect lives and landscapes from preventable loss, while enabling people to live in balance with nature.
This article was commissioned in conjunction with Prototypes for Humanity, a global initiative that showcases and accelerates academic innovation to solve social and environmental challenges. The Conversation is the media partner of Prototypes for Humanity 2025.