Unmanned Aerial Vehicles (UAVs) are improving considerably search and rescue (SAR) operations by providing unprecedented capabilities in dynamic and hazardous environments. This study presents an innovative, collaborative multi-UAV data fusion approach that addresses the critical challenge of locating multiple moving targets within strict time constraints. This approach improves traditional search techniques by incorporating intelligent information sharing, fusion, and coordinated path planning. The core innovation of the algorithm lies in its ability to dynamically and collaboratively predict the geographical zones with the highest probability of needed rescue operations. This enables the group of UAVs to coordinate and optimize their search strategies in real-time. This research offers valuable insights into multi-UAV collaboration through high-fidelity simulations involving more than 600 different scenarios with UAV swarms and moving ground targets. The experimental results indicate that their effectiveness significantly improves as the number of UAVs increases, following a quadratic trend until it reaches a plateau. In particular, the accuracy rate remains above 90%, regardless of the number of UAVs after reaching the plateau. This suggests that while a higher density of UAVs enhances search efficiency, larger UAV swarms yield diminishing returns. Notably, the approach shows superior efficiency in environments with clustered targets, which makes it particularly suitable for disaster response scenarios that involve more concentrated target locations.

The exponential growth of the Internet of Things (IoT) has introduced unprecedented challenges in data processing, privacy preservation, and energy efficiency. Traditional centralized approaches are often unsuitable for IoT environments due to bandwidth limitations, data heterogeneity, and privacy concerns. This study proposes a novel framework combining federated learning (FL) and autoencoders to address these issues in IoT-based image classification tasks. By lever-aging Unmanned Aerial Vehicles (UAVs) as intermediaries for model aggregation and distribution, the framework minimizes communication overhead while maintaining data privacy. Autoencoders are employed for unsupervised feature extraction, enabling effective data representation even in the absence of labeled data. Results demonstrate that, while autoencoders achieve lower classification accuracy compared to supervised approaches, they provide significant advantages in bandwidth efficiency, scalability, and privacy preservation. The integration of UAVs further enhances the system by optimizing communication and enabling model improvement in real-time. This framework offers a flexible and resource-efficient solution for IoT applications, particularly in scenarios where data labeling is impractical or privacy is paramount.

When dealing with large-scale natural disasters such as floods, landslides, hurricanes, or heavy snowfalls, there are often many victims who are trapped in hard-to-reach places, making the time to locate and rescue them critical. In this context, deploying unmanned aerial vehicles (UAVs) alongside a swarm of unmanned ground vehicles (UGVs) has the potential to speed up the Search and Rescue (SAR) missions, as the collaboration between these agents can combine advantages from both of them. The idea is that any simple UAV equipped GPS and communication capabilities can locate besieged and isolated individuals, referred to as Points of interest (POIs), and when it comes across (flies over) a GPS-limited UGV, it guides the UGVs toward the POIs. In this paper, we explore different approaches to Air-to-Ground (A2G) coordination in a number of distinct scenarios among unmanned vehicles, and through simulation compare their efficiency based of parameters and metrics.