There is a growing research field exploring how autonomous drones can enhance emergency response effectiveness. Integrating these (artificial) agents into existing emergency teams and workflows may significantly impact established accountability relationships. This paper examines how autonomous drones affect accountability attribution within complex socio-technical systems. Drawing on two real-life field trials in firefighting, the study reveals substantial uncertainty around accountability when drones are organizationally deployed. Using Bovens' accountability framework, two challenges are identified: (1) uncertainty about the role of drones within hierarchical structures, leading to confused accountability ascriptions; and (2) new forms of human-drone interactions introducing additional accountability-relevant issues. Based on these insights, the paper proposes actionable recommendations to support the responsible integration of autonomous drones into firefighting operations without undermining accountability. These findings offer practical guidance for policymakers and contribute to further research on accountability in autonomous systems.
This study presents the first systematic investigation of the dynamics of individual water droplets in the context of airtanker firefighting. While previous work has focused on ground-deposition patterns measured in standardized field tests, the droplet-scale mechanisms governing evaporation and transport have remained largely unexplored. A tailored model of the coupled momentum, heat, and mass transfer of an isolated water droplet in ambient air is proposed and applied to examine the evolution of droplets under a wide range of atmospheric conditions. The results demonstrate that droplet size governs the effectiveness of water delivery, the release height emerges as the dominant operational parameter, and relative humidity is the key atmospheric property. Increasing the release height lengthens the flight time and increases evaporative losses, while low relative humidity accelerates evaporation, particularly for droplets smaller than one millimeter. Only droplets within a narrow range of initial radii, $150\,μ\mathrm{m} \lesssim r_{\mathrm{d},0} \lesssim 3\,\mathrm{mm}$, are able to reach the ground following an airtanker release, with smaller droplets fully evaporating during thei
It is known that the online firefighting is 2-competitive on trees (Coupechoux et al. 2019), which suggests that the problem is relatively easy on trees. We extend the study to graphs containing cycles. We first show that the presence of cycles gives a strong advantage to the adversary: cycles create situations where the algorithm and the optimal solution operate on different game states, and the adversary can exploit the uncertainty in the firefighter sequence to trap the algorithm. Specifically, we prove that even on a tadpole graph (a cycle with a tail path), no deterministic online algorithm achieves a competitive ratio better than $Ω(\sqrt{n})$, where n is the number of vertices. We then propose an $O(\sqrt{n})$-competitive algorithm for 1-almost trees, which contain at most one cycle and generalize tadpole graphs. We further generalize this algorithm to cactus graphs, in which multiple cycles may appear, but no two share more than one vertex, and show that the online firefighting problem on cactus graphs remains $O(\sqrt{n})$-competitive. Finally, since cactus graphs have treewidth at most 2, we study a variant where firefighters are released in pairs, that is, each round an
Fires in industrial facilities pose special challenges to firefighters, e.g., due to the sheer size and scale of the buildings. The resulting visual obstructions impair firefighting accuracy, further compounded by inaccurate assessments of the fire's location. Such imprecision simultaneously increases the overall damage and prolongs the fire-brigades operation unnecessarily. We propose an automated assistance system for firefighting using a motorized fire monitor on a turntable ladder with aerial support from an unmanned aerial vehicle (UAV). The UAV flies autonomously within an obstacle-free flight funnel derived from geodata, detecting and localizing heat sources. An operator supervises the operation on a handheld controller and selects a fire target in reach. After the selection, the UAV automatically plans and traverses between two triangulation poses for continued fire localization. Simultaneously, our system steers the fire monitor to ensure the water jet reaches the detected heat source. In preliminary tests, our assistance system successfully localized multiple heat sources and directed a water jet towards the fires.
Inventory management of firefighting assets is crucial for emergency preparedness, risk assessment, and on-site fire response. However, conventional methods are inefficient due to limited capabilities in automated asset recognition and reconstruction. To address the challenge, this research introduces the Fire-ART dataset and develops a panoramic image-based reconstruction approach for semantic enrichment of firefighting assets into BIM models. The Fire-ART dataset covers 15 fundamental assets, comprising 2,626 images and 6,627 instances, making it an extensive and publicly accessible dataset for asset recognition. In addition, the reconstruction approach integrates modified cube-map conversion and radius-based spherical camera projection to enhance recognition and localization accuracy. Through validations with two real-world case studies, the proposed approach achieves F1-scores of 73% and 88% and localization errors of 0.620 and 0.428 meters, respectively. The Fire-ART dataset and the reconstruction approach offer valuable resources and robust technical solutions to enhance the accurate digital management of fire safety equipment.
Modern AI systems struggle most in environments where reliability is critical - scenes with smoke, poor visibility, and structural deformation. Each year, tens of thousands of firefighters are injured on duty, often due to breakdowns in situational perception. We introduce Fire360, a benchmark for evaluating perception and reasoning in safety-critical firefighting scenarios. The dataset includes 228 360-degree videos from professional training sessions under diverse conditions (e.g., low light, thermal distortion), annotated with action segments, object locations, and degradation metadata. Fire360 supports five tasks: Visual Question Answering, Temporal Action Captioning, Object Localization, Safety-Critical Reasoning, and Transformed Object Retrieval (TOR). TOR tests whether models can match pristine exemplars to fire-damaged counterparts in unpaired scenes, evaluating transformation-invariant recognition. While human experts achieve 83.5% on TOR, models like GPT-4o lag significantly, exposing failures in reasoning under degradation. By releasing Fire360 and its evaluation suite, we aim to advance models that not only see, but also remember, reason, and act under uncertainty. The
We consider a pursuit-evasion game that describes the process of extinguishing a fire burning on the nodes of an undirected graph. We denote the minimum number of firefighters required by ffn(G) and provide almost sharp bounds to this graph parameter for complete binary trees. We show that deciding whether ffn(G) <= m for given G and m is NP-hard. Furthermore, we show that shortest strategies can have superpolynomial length, leaving open whether the problem is in NP. We provide a construction that allows for transferring these results to a well-established Cops and Robbers variant called the "Hunter and Rabbit game".
As systems engineering (SE) objectives evolve from design and operation of monolithic systems to complex System of Systems (SoS), the discipline of Mission Engineering (ME) has emerged which is increasingly being accepted as a new line of thinking for the SE community. Moreover, mission environments are uncertain, dynamic, and mission outcomes are a direct function of how the mission assets will interact with this environment. This proves static architectures brittle and calls for analytically rigorous approaches for ME. To that end, this paper proposes an intelligent mission coordination methodology that integrates digital mission models with Reinforcement Learning (RL), that specifically addresses the need for adaptive task allocation and reconfiguration. More specifically, we are leveraging a Digital Engineering (DE) based infrastructure that is composed of a high-fidelity digital mission model and agent-based simulation; and then we formulate the mission tactics management problem as a Markov Decision Process (MDP), and employ an RL agent trained via Proximal Policy Optimization. By leveraging the simulation as a sandbox, we map the system states to actions, refining the policy
We present a forest fire firefighting simulation tool named FORFIS that is implemented in Python. Unlike other existing software, we focus on a user-friendly software interface with an easy-to-modify software engine. Our tool is published under GNU GPLv3 license and comes with a GUI as well as additional output functionality. The used wildfire model is based on the well-established approach by cellular automata in two variants - a rectangular and a hexagonal cell decomposition of the wildfire area. The model takes wind into account. In addition, our tool allows the user to easily include a customized firefighting strategy for the firefighting agents.
First responders risk their lives to reduce property damage and prevent injuries during disasters. Among first responders, firefighters work with fires in residential properties, forests, or other locations where fire occurs. We built the PyroGuardian system that uses wearable modules to transmit unit information over Long Range (LoRa) to an Android tablet. The tablet runs our application, PyroPortal, to assign each firefighter's stats, such as body temperature, heart rate, and GPS location. PyroPortal displays this information on unit dashboards, and markers on Google Maps represent the firefighter's location and the direction they are facing. These dashboards can help the incident commander (IC) make more informed decisions on mission control operations and remove specific units whose health stats, such as oximeter and pulse, passed certain thresholds. PyroGuardian completes all these tasks at an affordable cost and in an impressive maximum range between the units and IC. In addition, PyroGuardian has various application scenarios, such as law enforcement and military operations, besides firefighting. We also conducted a sample mission inside a burning building while real firefig
Firefighting is a complex, yet low automated task. To mitigate ergonomic and safety related risks on the human operators, robots could be deployed in a collaborative approach. To allow human-robot teams in firefighting, important basics are missing. Amongst other aspects, the robot must predict the human motion as occlusion is ever-present. In this work, we propose a novel motion prediction pipeline for firefighters' squads in indoor search and rescue. The squad paths are generated with an optimal graph-based planning approach representing firefighters' tactics. Paths are generated per room which allows to dynamically adapt the path locally without global re-planning. The motion of singular agents is simulated using a modification of the headed social force model. We evaluate the pipeline for feasibility with a novel data set generated from real footage and show the computational efficiency.
During a wildfire, the work of the aerial coordinator is crucial for the control of the wildfire and the minimization of the burned area and the damage caused. Since it could be very useful for the coordinator to have decision-making tools at his/her disposal, this framework deals with an optimization model to obtain the optimal planning of firefighting helicopters, deciding the points where the aircraft should load water, the areas of the wildfire where they should work, and the rest bases to which each helicopter should be assigned. It was developed a Mixed Integer Linear Programming model which takes into account the configuration of helicopters, in closed circuits, as well as the flight aerial regulations in Spain. Due to the complexity of the model, two algorithms are developed, based on the Simulated Annealing and Iterated Local Search metaheuristic techniques. Both algorithms are tested with real data instances, obtaining very promising results for future application in the planning of aircraft throughout a wildfire evolution.
In the classic version of the game of firefighter, on the first turn a fire breaks out on a vertex in a graph $G$ and then $b$ firefighters protect $b$ vertices. On each subsequent turn, the fire spreads to the collective unburned neighbourhood of all the burning vertices and the firefighters again protect $b$ vertices. Once a vertex has been burned or protected it remains that way for the rest of the game. In \textit{distance-restricted firefighting} the firefighters' movement is restricted so they can only move up to some fixed distance $d$ and they may or may not be permitted to move through burning vertices. In this paper we establish the NP-completeness of the distance-restricted versions of {\sc $b$-Firefighter} and present an integer program for computing the exact value. We also discuss some interesting properties of the \textit{Expected Damage} function.
The Firefighting problem is defined as follows. At time $t=0$, a fire breaks out at a vertex of a graph. At each time step $t \geq 0$, a firefighter permanently defends (protects) an unburned vertex, and the fire then spread to all undefended neighbors from the vertices on fire. This process stops when the fire cannot spread anymore. The goal is to find a sequence of vertices for the firefighter that maximizes the number of saved (non burned) vertices. The Firefighting problem turns out to be NP-hard even when restricted to bipartite graphs or trees of maximum degree three. We study the parameterized complexity of the Firefighting problem for various structural parameterizations. All our parameters measure the distance to a graph class (in terms of vertex deletion) on which the firefighting problem admits a polynomial time algorithm. Specifically, for a graph class $\mathcal{F}$ and a graph $G$, a vertex subset $S$ is called a modulator to $\mathcal{F}$ if $G \setminus S$ belongs to $\mathcal{F}$. The parameters we consider are the sizes of modulators to graph classes such as threshold graphs, bounded diameter graphs, disjoint unions of stars, and split graphs. To begin with, we sh
Continuous and discrete models for firefighting problems are well-studied in Theoretical Computer Science. We introduce a new, discrete, and more general framework based on a hexagonal cell graph to study firefighting problems in varied terrains. We present three different firefighting problems in the context of this model; for two of which, we provide efficient polynomial time algorithms and for the third, we show NP-completeness. We also discuss possible extensions of the model and their implications on the computational complexity.
In the classic version of the game of firefighter, on the first turn a fire breaks out on a vertex in a graph $G$ and then $k$ firefighters protect $k$ vertices. On each subsequent turn, the fire spreads to the collective unburnt neighbourhood of all the burning vertices and the firefighters again protect $k$ vertices. Once a vertex has been burnt or protected it remains that way for the rest of the game. A common objective with respect to some infinite graph $G$ is to determine how many firefighters are necessary to stop the fire from spreading after a finite number of turns, commonly referred to as containing the fire. We introduce the concept of distance-restricted firefighting where the firefighters' movement is restricted so they can only move up to some fixed distance $d$ per turn rather than being able to move without restriction. We establish some general properties of this new game in contrast to properties of the original game, and we investigate specific cases of the distance-restricted game on the infinite square, strong, and hexagonal grids. We conjecture that two firefighters are insufficient on the square grid when $d = 2$, and we pose some questions about how many f
We study Hartnell's firefighter problem on infinite trees and characterise the branching number in terms of the firefighting game. Using our results about trees, we give a partial answer to a question of Martínez-Pedroza concerning firefighting on Cayley graphs.
Drone swarms coupled with data intelligence can be the future of wildfire fighting. However, drone swarm firefighting faces enormous challenges, such as the highly complex environmental conditions in wildfire scenes, the highly dynamic nature of wildfire spread, and the significant computational complexity of drone swarm operations. We develop a predict-then-optimize approach to address these challenges to enable effective drone swarm firefighting. First, we construct wildfire spread prediction convex neural network (Convex-NN) models based on real wildfire data. Then, we propose a mixed-integer programming (MIP) model coupled with dynamic programming (DP) to enable efficient drone swarm task planning. We further use chance-constrained robust optimization (CCRO) to ensure robust firefighting performances under varying situations. The formulated model is solved efficiently using Benders Decomposition and Branch-and-Cut algorithms. After 75 simulated wildfire environments training, the MIP+CCRO approach shows the best performance among several testing sets, reducing movements by 37.3\% compared to the plain MIP. It also significantly outperformed the GA baseline, which often failed t
Allocation of personnel and material resources is highly sensible in the case of firefighter interventions. This allocation relies on simulations to experiment with various scenarios. The main objective of this allocation is the global optimization of the firefighters response. Data generation is then mandatory to study various scenarios In this study, we propose to compare different data generation methods. Methods such as Random Sampling, Tabular Variational Autoencoders, standard Generative Adversarial Networks, Conditional Tabular Generative Adversarial Networks and Diffusion Probabilistic Models are examined to ascertain their efficacy in capturing the intricacies of firefighter interventions. Traditional evaluation metrics often fall short in capturing the nuanced requirements of synthetic datasets for real-world scenarios. To address this gap, an evaluation of synthetic data quality is conducted using a combination of domain-specific metrics tailored to the firefighting domain and standard measures such as the Wasserstein distance. Domain-specific metrics include response time distribution, spatial-temporal distribution of interventions, and accidents representation. These m
In firefighting and other emergency operations, decisions made under pressure carry profound ethical weight and can significantly impact incident outcomes and firefighter safety. Traditional training methods, while foundational, often fall short in adequately preparing firefighters for the complex ethical dilemmas and value conflicts inherent in chaotic emergency environments. This paper proposes a conceptual framework for enhancing firefighter training by systematically integrating departmental values into simulation-based training. This approach fosters deeper value internalisation and improves value-driven decision-making under pressure. Furthermore, the underlying tools can also be leveraged to evaluate and refine departmental operational protocols for better alignment with preferred values.