In recent years, the use of multiple unmanned aerial vehicles (UAVs) in various applications has progressively increased thanks to advancements in multi-agent system technology, which enables the accomplishment of complex tasks that require cooperative and coordinated abilities. In this article, multi-UAV applications are grouped into five classes based on their primary task: coverage, adversarial search and game, computational offloading, communication, and target-driven navigation. By employing a systematic review approach, we select the most significant works that use deep reinforcement learning (DRL) techniques for cooperative and scalable multi-UAV systems and discuss their features using extensive and constructive critical reasoning. Finally, we present the most likely and promising research directions by highlighting the limitations of the currently held assumptions and the constraints when dealing with collaborative DRL-based multi-UAV systems. The suggested areas of research can enhance the transfer of knowledge from simulations to real-world environments and can increase the responsiveness and safety of UAV systems.

The interest in applications related to Multi-Unmanned Aerial Vehicle (UAV) systems has been growing exponentially in the last few years. Reinforcement Learning (RL) presents one of the most popular alternatives for solving Multi-UAV tasks, thanks to its flexible requirements for modeling the problem. However, it is often applied to abstractions of the original problem, thus leaving to next development phases the integration of RL solutions to the actual systems. This choice may not guarantee the overall optimal performance of the implemented system. In this survey, we analyze the literature on Multi-UAV applications that utilize reinforcement learning, with particular attention to works that consider realistic communication channels. We focus on identifying the key variables that influence communication and whether these variables are integrated within the RL framework or considered externally. Additionally, we identify key trends, challenges, and future directions in the field, providing a comprehensive overview for researchers and practitioners interested in the practical deployment of RL-based Multi-UAV systems.

Human-AI teamwork is no longer a topic of the future. Given the importance of trust in human teams, the question arises how trust functions in human-AI teams. Although trust has long been studied from a human-centred perspective (e.g. in psychology and philosophy), a computational perspective and from the perspective of human trust in AI (e.g. in human-computer interaction), the study of trust in human-AI interaction in a team setting is still a novel field. For this reason, the MULTITTRUST (Multidisciplinary perspectives on Human-AI Team Trust) workshop series was founded. In this paper, we present the main outcomes after three editions. Our contributions are: an overview of the shared language of concepts and definitions; an outline of the main open research challenges; and methodological guidelines for further studies in meaningful human-AI team trust. These three contributions form a foundational roadmap towards a better understanding of trust in human-AI team interactions.

Trust is a well-established and extensively studied concept in the humanitarian field, yet its highly subjective nature, dependence on a wide range of influencing factors, and the absence of a single, unified definition pose significant challenges when attempting to apply it to Human-AI teams. However, being able to assess and recognize when to trust an agent is crucial, especially when deciding whether to delegate tasks or to collaborate on more intricate tasks within a team. Existing work has studied the integration of specific computable factors, which are known to influence trust. In this work, we study how these methods transfer to a more concrete Human-AI collaborative puzzle game. In this setting, the AI agent employs a trust-aware Reinforcement Learning algorithm, while the Human agent’s behavior is emulated through the use of a Large Language Model. Our study considers the realistic condition of heterogeneous agents, each skilled in different areas, and we demonstrate that trust-awareness, particularly the ability to make justified decisions about task delegation, is essential for the successful operation of the team. This scenario serves as a testbed for evaluating methods of trust integration and their impact on Human-AI collaboration.

The use of Unmanned Aerial Vehicle (UAVs) in deep reinforcement learning (DRL) applications is growing rapidly, as DRL is effective in handling complex environments. One major challenge is designing effective reward signals, as sparse rewards hinder the learning process of DRL. On the other hand, tabular reinforcement learning (RL) approaches, while faster, are typically limited to small, discrete environments and do not scale well to more complex settings. This work presents a hierarchical approach combining the strengths of both methods. We apply tabular RL to derive an optimal value function for a discrete abstract version of the original environment, and use it as a heuristic to generate a denser reward signal for the DRL algorithm, in the original environment. We empirically demonstrate that our approach significantly reduces the number of samples the DRL algorithm needs to achieve the same level of performance in a multi-UAV scenario, compared to traditional methods.