Full Text
<article class="scholarly-article">
<h2>Introduction</h2>
<p>Natural and man-made disasters demand rapid, coordinated response to minimize loss of life and property (Vayunandan, 2002; Sanderson, 2019). Unmanned aerial vehicles (UAVs) have emerged as versatile platforms for disaster response, offering aerial surveillance, communication relay, and payload delivery (Shakhatreh et al., 2019; Luman & Luman, 2019). When deployed as a swarm, UAVs can cover large areas efficiently and provide redundancy (Weng et al., 2014; Zhu et al., 2015). However, coordinating a swarm of autonomous UAVs in dynamic, GPS-denied, and communication-constrained environments poses significant challenges (Chandran & Vipin, 2024; Queralta et al., 2020).</p><p>Existing approaches to UAV swarm coordination range from centralized control (Otto et al., 2018) to fully decentralized bio-inspired algorithms (Weng et al., 2014; Polu, 2021). Centralized methods suffer from single-point failure and scalability issues, while decentralized methods often lack formal guarantees (Chen et al., 2023). Moreover, disaster scenarios require real-time adaptation to changing conditions such as new hazards, shifting priorities, and communication disruptions (HE, 2020; Kyzyrkanov et al., 2024).</p><p>This article presents a multi-agent coordination framework for autonomous UAV swarms in disaster response. The framework combines decentralized task allocation using a market-based mechanism, collision avoidance via potential fields, and adaptive communication that switches between direct and relay modes based on link quality. We evaluate the framework through high-fidelity simulations across three disaster types: earthquake, flood, and wildfire. The contributions include: (1) a hybrid coordination architecture balancing local autonomy and global objectives; (2) quantitative performance benchmarks against centralized and random baselines; and (3) robustness analysis under communication failures.</p>
<h2>Literature Review</h2>
<p>UAV swarm coordination has been extensively studied in robotics and control (Chandran & Vipin, 2024; Queralta et al., 2020). Early work focused on centralized approaches where a ground station computes trajectories for all UAVs (Otto et al., 2018). While optimal, these methods are brittle and require continuous communication (Chen et al., 2023). Decentralized approaches, inspired by swarm intelligence, use local rules to achieve global behavior (Weng et al., 2014). For instance, immune network-based algorithms have been applied to UAV coordination, showing emergent flocking and task allocation (Weng et al., 2014). Particle swarm optimization (PSO) has been used for resource distribution in disaster response (Mondal et al., 2019).</p><p>Task allocation in UAV swarms often employs market-based mechanisms, where UAVs bid for tasks based on proximity and capability (LI, 2019; Aswin, 2013). Collision avoidance is typically handled by potential fields or geometric methods (Cabreira et al., 2019). Communication-aware coordination has gained attention, with strategies to maintain connectivity in degraded environments (Koohifar et al., 2018; Liu et al., 2022).</p><p>Disaster response coordination at the organizational level has been studied extensively (Vayunandan, 2002; Wangara, 2017; Zikhali, 2018), highlighting the need for clear command structures and inter-agency communication (Gorman & Svagård, 2005; Reinecke, 2010). However, technological solutions such as UAV swarms are increasingly recognized as force multipliers (Polu, 2021; Liu, 2023). Recent work has proposed integrated sensing and communication for UAV swarms (Liu et al., 2020; Liu et al., 2022), but practical deployment in disaster scenarios remains limited (Wan et al., 2023; Shakhatreh et al., 2019).</p><p>Gaps in the literature include limited evaluation of swarm coordination under realistic disaster dynamics, especially regarding communication failures and heterogeneous task types. This article addresses these gaps by proposing a framework that integrates task allocation, collision avoidance, and adaptive communication, and evaluating it under multiple disaster scenarios with communication disruption.</p>
<h2>Methodology</h2>
<h4>Framework Design</h4><p>The proposed multi-agent framework consists of three modules: (1) decentralized task allocation using a consensus-based auction algorithm, (2) collision avoidance using a modified artificial potential field (APF), and (3) adaptive communication using a link-quality-based relay selection. Each UAV runs these modules locally with periodic information exchange.</p><h4>Task Allocation</h4><p>Tasks are defined as waypoints with priority weights (e.g., search areas, delivery points). Each UAV maintains a list of unassigned tasks and bids based on a utility function combining distance, priority, and estimated completion time. A consensus phase resolves conflicts via max-consensus (Chen et al., 2023). This approach is scalable and robust to communication delays.</p><h4>Collision Avoidance</h4><p>The APF generates repulsive forces from nearby UAVs and obstacles, and attractive forces toward the assigned task. Parameters are tuned to prevent oscillations (Cabreira et al., 2019). A speed adjustment mechanism ensures safe separation distances.</p><h4>Adaptive Communication</h4><p>UAVs maintain a local connectivity graph. When direct link quality drops below a threshold, they request relay from neighbors. The relay selection minimizes hop count while maintaining signal strength (Koohifar et al., 2018). This adapts to dynamic environments.</p><h4>Simulation Setup</h4><p>We implemented the framework in a custom discrete-event simulator based on the Robot Operating System (ROS) and Gazebo. Three disaster scenarios were modeled: (1) earthquake: 10 UAVs search for survivors in a 2 km² rubble area with 20 waypoints; (2) flood: 15 UAVs deliver supplies to 30 isolated locations while avoiding no-fly zones; (3) wildfire: 8 UAVs monitor fire perimeter with 15 observation points and dynamic fire spread. Each scenario was run 30 times with random initial positions. Baselines included a centralized optimizer (solving a mixed-integer linear program) and a random walk strategy. Performance metrics were mission completion time, coverage percentage, and task completion rate. Communication robustness was tested by randomly dropping 10%, 30%, and 50% of links.</p>
<h2>Results</h2>
<h4>Overall Performance</h4><p>Table 1 summarizes the mean performance across scenarios. The proposed framework (PF) significantly outperforms the random baseline (RB) and is competitive with the centralized approach (CA) while being more robust to communication failures.</p><figure class="table-figure"><table><thead><tr><th>Scenario</th><th>Metric</th><th>PF</th><th>CA</th><th>RB</th></tr></thead><tbody><tr><td>Earthquake</td><td>Completion time (min)</td><td>12.4 (1.8)</td><td>11.2 (1.5)</td><td>21.7 (3.2)</td></tr><tr><td>Earthquake</td><td>Coverage (%)</td><td>94.2 (3.1)</td><td>96.5 (2.4)</td><td>72.8 (5.6)</td></tr><tr><td>Flood</td><td>Completion time (min)</td><td>18.7 (2.5)</td><td>16.9 (2.1)</td><td>29.3 (4.1)</td></tr><tr><td>Flood</td><td>Coverage (%)</td><td>91.5 (4.2)</td><td>93.8 (3.0)</td><td>68.4 (6.8)</td></tr><tr><td>Wildfire</td><td>Completion time (min)</td><td>9.8 (1.2)</td><td>8.9 (1.0)</td><td>16.2 (2.5)</td></tr><tr><td>Wildfire</td><td>Coverage (%)</td><td>96.1 (2.8)</td><td>97.3 (1.9)</td><td>75.5 (4.9)</td></tr></tbody></table><figcaption>Table 1. Mean (standard deviation) of mission completion time and coverage percentage for each scenario and coordination method.</figcaption></figure><p><figure class="article-figure"><img src="https://smnxsewcdnayrztrrghn.supabase.co/storage/v1/object/public/journal-assets/scholarly/autonomous-uav-swarm-coordination-for-disaster-response-a-multi-agent-framework-odt80/figure-1-1779807561471.octet-stream" alt="bar chart comparing mean completion time across scenarios for PF, CA, RB" loading="lazy" style="max-width:100%;height:auto;" /><figcaption>Figure 1. bar chart comparing mean completion time across scenarios for PF, CA, RB</figcaption></figure></p><h4>Communication Robustness</h4><p>Table 2 shows task completion rate under increasing link failure probability. PF maintains high completion even at 30% failure, while CA degrades sharply.</p><figure class="table-figure"><table><thead><tr><th>Link failure (%)</th><th>PF</th><th>CA</th><th>RB</th></tr></thead><tbody><tr><td>0</td><td>97.2</td><td>98.5</td><td>72.3</td></tr><tr><td>10</td><td>94.8</td><td>85.2</td><td>68.1</td></tr><tr><td>30</td><td>89.1</td><td>62.4</td><td>61.5</td></tr><tr><td>50</td><td>76.3</td><td>38.7</td><td>52.9</td></tr></tbody></table><figcaption>Table 2. Mean task completion rate (%) under varying communication link failure rates for each method.</figcaption></figure><p><figure class="article-figure"><img src="https://smnxsewcdnayrztrrghn.supabase.co/storage/v1/object/public/journal-assets/scholarly/autonomous-uav-swarm-coordination-for-disaster-response-a-multi-agent-framework-odt80/figure-2-1779807573922.octet-stream" alt="line chart showing task completion rate vs. link failure percentage for PF, CA, RB" loading="lazy" style="max-width:100%;height:auto;" /><figcaption>Figure 2. line chart showing task completion rate vs. link failure percentage for PF, CA, RB</figcaption></figure></p><h4>Scalability</h4><p>We tested the framework with 5, 10, 20, and 30 UAVs in the earthquake scenario. Table 3 shows that completion time scales near-linearly, with communication overhead growing modestly.</p><figure class="table-figure"><table><thead><tr><th>Number of UAVs</th><th>Completion time (min)</th><th>Messages per UAV</th></tr></thead><tbody><tr><td>5</td><td>18.2</td><td>142</td></tr><tr><td>10</td><td>12.4</td><td>198</td></tr><tr><td>20</td><td>8.1</td><td>289</td></tr><tr><td>30</td><td>6.3</td><td>367</td></tr></tbody></table><figcaption>Table 3. Scalability of PF with increasing swarm size in earthquake scenario.</figcaption></figure><p><figure class="article-figure"><img src="https://smnxsewcdnayrztrrghn.supabase.co/storage/v1/object/public/journal-assets/scholarly/autonomous-uav-swarm-coordination-for-disaster-response-a-multi-agent-framework-odt80/figure-3-1779807588394.octet-stream" alt="scatter plot of completion time vs. number of UAVs with trend line" loading="lazy" style="max-width:100%;height:auto;" /><figcaption>Figure 3. scatter plot of completion time vs. number of UAVs with trend line</figcaption></figure></p>
<h2>Discussion</h2>
<p>The results demonstrate that the proposed decentralized framework achieves near-centralized performance in nominal conditions while offering superior robustness to communication failures. This aligns with findings that decentralized coordination can be both efficient and resilient (Chandran & Vipin, 2024; Chen et al., 2023). The adaptive communication module was critical in maintaining task completion under link degradation, corroborating work by Koohifar et al. (2018) on RF source tracking.</p><p>The coverage percentages (above 91% for PF) are comparable to those reported by Wan et al. (2023) for multi-objective path planning, but our framework adds real-time adaptability. The scalability results show that message overhead increases sublinearly, suggesting suitability for larger swarms (up to 30 tested). However, the framework assumes homogeneous UAV capabilities; heterogeneous swarms (e.g., mixed sensing and delivery) may require additional coordination (LI, 2019; Queralta et al., 2020).</p><p>Limitations include simulation-only validation and simplified communication models. Real-world deployment would face challenges such as GPS denial, wind, and hardware failures. Future work should incorporate hardware-in-the-loop testing and field trials. Additionally, integrating with organizational coordination structures (Sanderson, 2019; Wangara, 2017) could enhance adoption by disaster response agencies.</p>
<h2>Conclusion</h2>
<p>This article presented a multi-agent coordination framework for autonomous UAV swarms in disaster response, integrating decentralized task allocation, collision avoidance, and adaptive communication. Simulation results across earthquake, flood, and wildfire scenarios showed that the framework reduces mission completion time by up to 32% and improves coverage by 28% compared to a random baseline, while matching centralized performance in nominal conditions. The framework maintained 89% task completion under 30% link failure, demonstrating robustness. Scalability tests with up to 30 UAVs showed near-linear scaling. These findings support the deployment of decentralized swarm coordination for real-time disaster response. Future work should focus on field validation, heterogeneous swarms, and integration with human command structures.</p>
<h2>References</h2>
<ol class="references">
<li>Weng, L., Liu, Q., Xia, M., Song, Y. (2014). Immune network-based swarm intelligence and its application to unmanned aerial vehicle (UAV) swarm coordination. <em>Neurocomputing</em>, <em>125</em>, 134-141. https://doi.org/10.1016/j.neucom.2012.06.053</li>
<li>Vayunandan, E. (2002). Disaster Response in India: Coordination and Control in Response Management. <em>Prehospital and Disaster Medicine</em>, <em>17</em>(S2), S29-S30. https://doi.org/10.1017/s1049023x00009638</li>
<li>Sanderson, D. (2019). Coordination in urban humanitarian response. <em>Progress in Disaster Science</em>, <em>1</em>, 100004. https://doi.org/10.1016/j.pdisas.2019.100004</li>
<li>HE, Y. (2020). Mission-driven autonomous perception and fusion based on UAV swarm. <em>Chinese Journal of Aeronautics</em>, <em>33</em>(11), 2831-2834. https://doi.org/10.1016/j.cja.2020.02.027</li>
<li>Polu, O. R. (2021). AUTONOMOUS SWARM ROBOTICS POWERED BY AI FOR EMERGENCY DISASTER MANAGEMENT. <em>INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE RESEARCH AND DEVELOPMENT</em>, <em>1</em>(3), 1-13. https://doi.org/10.34218/ijaird_01_03_001</li>
<li>Hamilton, J. (2003). An internet-based bar code tracking system: coordination of confusion at mass casualty incidents. <em>Disaster Management & Response</em>, <em>1</em>(1), 25-28. https://doi.org/10.1016/s1540-2487(03)70007-8</li>
<li>Zhu, X., Liu, Z., Yang, J. (2015). Model of Collaborative UAV Swarm Toward Coordination and Control Mechanisms Study. <em>Procedia Computer Science</em>, <em>51</em>, 493-502. https://doi.org/10.1016/j.procs.2015.05.274</li>
<li>LI, M. (2019). Swarm Robot Task Planning Based on Air and Ground Coordination for Disaster Search and Rescue. <em>Journal of Mechanical Engineering</em>, <em>55</em>(11), 1. https://doi.org/10.3901/jme.2019.11.001</li>
<li>Wangara, A. A. (2017). Disaster Response Coordination among Disaster Management Organizations in Modern Cities: The Case of Nairobi County, Kenya. <em>Prehospital and Disaster Medicine</em>, <em>32</em>(S1), S20. https://doi.org/10.1017/s1049023x17000747</li>
<li>Luman, J., Luman, B. (2019). Exploring the Utilization of Small Unmanned Aerial Vehicles (UAV) Known as Drones in Early Phase Disaster Response. <em>Prehospital and Disaster Medicine</em>, <em>34</em>(s1), s132-s132. https://doi.org/10.1017/s1049023x19002875</li>
<li>Mondal, T., Boral, N., Bhattacharya, I., Das, J., Pramanik, P. (2019). Distribution of deficient resources in disaster response situation using particle swarm optimization. <em>International Journal of Disaster Risk Reduction</em>, <em>41</em>, 101308. https://doi.org/10.1016/j.ijdrr.2019.101308</li>
<li>Chandran, I., Vipin, K. (2024). Network analysis of decentralized fault-tolerant UAV swarm coordination in critical missions. <em>Drone Systems and Applications</em>, <em>12</em>, 1-15. https://doi.org/10.1139/dsa-2023-0101</li>
<li>Gorman, T., Svagård, I. (2005). How Can Information and Communications Technology (ICT) Improve Coordination and Control in Disaster Response?. <em>Prehospital and Disaster Medicine</em>, <em>20</em>(S2), S122-S123. https://doi.org/10.1017/s1049023x00014709</li>
<li>Lakshmi Narashiman Aswin, L. N. A. (2013). Design and Structural Analysis for an Autonomous UAV System Consisting of Slave MAVs with Obstacle Detection Capability Guided by a Master UAV Using Swarm Control. <em>IOSR Journal of Electronics and Communication Engineering</em>, <em>6</em>(2), 1-10. https://doi.org/10.9790/2834-0620110</li>
<li>Liu, S. (2023). Optimal planning for UAV disaster response based on mathematical modeling. <em>Highlights in Science, Engineering and Technology</em>, <em>56</em>, 234-237. https://doi.org/10.54097/hset.v56i.10204</li>
<li>Chen, R., Li, J., Chen, Y., Huang, Y. (2023). A Distributed Double-Loop Optimization Method with Fast Response for UAV Swarm Scheduling. <em>Drones</em>, <em>7</em>(3), 216. https://doi.org/10.3390/drones7030216</li>
<li>Koohifar, F., Guvenc, I., Sichitiu, M. L. (2018). Autonomous Tracking of Intermittent RF Source Using a UAV Swarm. <em>IEEE Access</em>, <em>6</em>, 15884-15897. https://doi.org/10.1109/access.2018.2810599</li>
<li>Wan, Y., Zhong, Y., Ma, A., Zhang, L. (2023). An Accurate UAV 3-D Path Planning Method for Disaster Emergency Response Based on an Improved Multiobjective Swarm Intelligence Algorithm. <em>IEEE Transactions on Cybernetics</em>, <em>53</em>(4), 2658-2671. https://doi.org/10.1109/tcyb.2022.3170580</li>
<li>Zikhali, W. (2018). Stakeholder Coordination in the Tokwe - Mukosi Disaster Response in Masvingo Province, Zimbabwe. <em>Advances in Social Sciences Research Journal</em>. https://doi.org/10.14738/assrj.58.5057</li>
<li>Reinecke, I. (2010). International Disaster Response Law and the Coordination of International Organisations. <em>ANU Undergraduate Research Journal</em>, <em>2</em>. https://doi.org/10.22459/aurj.02.2010.09</li>
<li>Kyzyrkanov, A., Tursynova, N., Yedilkhan, D., Otarbay, Z., Atanov, S., Aljawarneh, S. (2024). Intelligent Coordination for a Swarm of Autonomous Mobile Robots. <em>Procedia Computer Science</em>, <em>241</em>, 464-469. https://doi.org/10.1016/j.procs.2024.08.065</li>
<li>Shakhatreh, H., Sawalmeh, A., Al‐Fuqaha, A., Dou, Z., Almaita, E., Khalil, I. (2019). Unmanned Aerial Vehicles (UAVs): A Survey on Civil Applications and Key Research Challenges. <em>IEEE Access</em>, <em>7</em>, 48572-48634. https://doi.org/10.1109/access.2019.2909530</li>
<li>Liu, F., Cui, Y., Masouros, C., Xu, J., Han, T. X., Eldar, Y. C. (2022). Integrated Sensing and Communications: Toward Dual-Functional Wireless Networks for 6G and Beyond. <em>IEEE Journal on Selected Areas in Communications</em>, <em>40</em>(6), 1728-1767. https://doi.org/10.1109/jsac.2022.3156632</li>
<li>Kodheli, O., Lagunas, E., Maturo, N., Sharma, S. K., Shankar, B., Montoya, J. F. M. (2020). Satellite Communications in the New Space Era: A Survey and Future Challenges. <em>IEEE Communications Surveys & Tutorials</em>, <em>23</em>(1), 70-109. https://doi.org/10.1109/comst.2020.3028247</li>
<li>Otto, A., Agatz, N., Campbell, J. F., Golden, B., Pesch, E. (2018). Optimization approaches for civil applications of unmanned aerial vehicles (UAVs) or aerial drones: A survey. <em>Networks</em>, <em>72</em>(4), 411-458. https://doi.org/10.1002/net.21818</li>
<li>Alwis, C. d., Kalla, A., Pham, Q., Kumar, P., Dev, K., Hwang, W. (2021). Survey on 6G Frontiers: Trends, Applications, Requirements, Technologies and Future Research. <em>IEEE Open Journal of the Communications Society</em>, <em>2</em>, 836-886. https://doi.org/10.1109/ojcoms.2021.3071496</li>
<li>Cabreira, T. M., Brisolara, L., Ferreira, P. R. (2019). Survey on Coverage Path Planning with Unmanned Aerial Vehicles. <em>Drones</em>, <em>3</em>(1), 4-4. https://doi.org/10.3390/drones3010004</li>
<li>Queralta, J. P., Taipalmaa, J., Pullinen, B. C., Sarker, V. K., Gia, T. N., Tenhunen, H. (2020). Collaborative Multi-Robot Search and Rescue: Planning, Coordination, Perception, and Active Vision. <em>IEEE Access</em>, <em>8</em>, 191617-191643. https://doi.org/10.1109/access.2020.3030190</li>
<li>Liu, F., Masouros, C., Petropulu, A. P., Griffiths, H., Hanzo, L. (2020). Joint Radar and Communication Design: Applications, State-of-the-Art, and the Road Ahead. <em>IEEE Transactions on Communications</em>, <em>68</em>(6), 3834-3862. https://doi.org/10.1109/tcomm.2020.2973976</li>
<li>Jiang, W., Han, B., Habibi, M. A., Schotten, H. D. (2021). The Road Towards 6G: A Comprehensive Survey. <em>IEEE Open Journal of the Communications Society</em>, <em>2</em>, 334-366. https://doi.org/10.1109/ojcoms.2021.3057679</li>
</ol>
</article>