Overcoming UAV Limitations Through Swarm Intelligence
Abstract: Although UAV technology has profoundly transformed the core operations of numerous industries, critical challenges persist, such as limited battery life, excessive human dependence, and the pressing need for real-time adaptive mission planning. These issues significantly hinder the application of UAV technologies in complex, computationally demanding, and time-consuming missions, ultimately constraining their full potential. While future advancements in battery systems, hardware, and software will mitigate these problems, accelerating the development of swarming technology can exponentially increase adoption and unlock a completely new set of applications that are currently beyond consideration. AURA Lab attempts to alleviate these limitations by developing algorithms and systems for the efficient, scalable deployment of multiple UAVs. Naturally, this work is organized around three pillars: i) Offline planning, ii) Online learning and adaptation, and iii) Dataset acquisition and construction. In offline planning, AURA Lab focuses on extracting the maximum value from available pre-mission information, designing methods for balanced multi-UAV coverage, optimal swarm deployment, and efficient inspection of both large areas and scattered targets. In online learning, the goal is to enable UAV swarms to adapt in real time to dynamic environments, changing swarm configurations, and evolving mission demands, making them more resilient and effective in tasks such as surveillance, precision agriculture, and gas-plume tracking. Finally, AURA Lab develops and maintains open UAV datasets to lower the barrier for further research and innovation in the drone ecosystem.
Bio: Athanasios Kapoutsis is an Assistant Professor in the Department of Electrical and Computer Engineering (ECE) at Democritus University of Thrace (DUTH) and the head of the AUtonomous Robotics Agents (AURA) Lab. His research interests include, i) inducing autonomy in real-world multi-robot systems, ii) spanning coverage/exploration path planning for UAVs/UGVs/AUVs, iii) reinforcement learning and multi-agent RL, and iv) large-scale distributed/adaptive control. During the past years, he has contributed to a broad portfolio of competitively funded R&D projects spanning EU FP7, H2020 and Horizon Europe programmes, as well as Greek national initiatives—taking on responsibilities ranging from key technical roles to project leadership. Across his work in competitively funded R&D, he has served through three institutional affiliations: the Information Technologies Institute (ITI) of the Centre for Research & Technology Hellas (CERTH), the Department of Electrical and Computer Engineering (ECE) at the Democritus University of Thrace (DUTH), and the Institute for Language and Speech Processing (ILSP) of the Athena Research and Innovation Center (ARC).