Driving behavior analysis is important for road safety and improving traffic efficiency. Knowledge of how driving styles differ can help policymakers and urban planners design interventions that reduce accidents and improve general traffic flow. With the advent of Unmanned Aerial Vehicles (UAVs), the scope and accuracy of information retrieval have significantly increased, mainly due to benefits such as greater coverage, flexibility, and cost efficiency compared to traditional sensors such as GPS. The goal of this work is to classify different driving behaviors, predominantly using some machine learning models, on the huge pNEUMA dataset constructed through measurements of natural vehicle trajectory by a swarm of UAVs over Athens, Greece. Our analysis distinguishes between cautious, normal, and aggressive driving behaviors across different vehicle types. New relevant features are extracted to further increase the robustness of our model in the parameters of the original dataset. Noticeably, the Kolmogorov-Arnold Network (KAN) model reached higher performance with 99% accuracy. All of these—high accuracy and huge potential for KAN in capturing complex driving patterns and their ability to adapt to dynamic traffic environments—stress their promise. Such findings could be used to help develop the next generation of traffic management systems and understand driving behaviors at an even deeper level.
The adoption of novel technologies such as Unmanned Aerial Vehicles (UAVs) in Search and Rescue (SAR) operations remains limited. As a result, their full potential is not yet realized. Although UAVs have been deployed on an ad hoc basis, typically under manual control by dedicated operators, assisted and fully autonomous configurations remain largely unexplored. In this study, three SAR frameworks are systematically evaluated within a unified benchmarking framework: conventional ground missions, UAV-assisted missions, and fully autonomous UAV operations. As the key performance indicator, the target localization time was quantified and used as the means of comparison amongst frameworks. The conventional and assisted frameworks were experimentally tested through physical hardware in a controlled outdoor setting, wherein simulated callouts occurred via rescue teams. The autonomous swarm framework was simulated in the form of a multi-agent Reinforcement Learning (RL) method via the use of the Proximal Policy Optimization (PPO) algorithm. This enabled the optimization of the decentralized cooperative actions that could occur for efficient exploration of a partially observed three-dimensional environment. Our results demonstrated that the autonomous swarm significantly outperformed the conventional and assisted approaches in terms of speed and coverage. Finally, a detailed depiction of the framework’s integration into an operational system is provided.