An Electrified Multi-UAV-Based State Estimation for High-Precision Perception Tasks

The increasing demand for high-precision surveillance has driven interest in lightweight alternatives to traditional ground-based systems. Recent advancements in the electrification of aerial platforms have facilitated the development and deployment of various types of electric unmanned aerial vehicles (UAVs), contributing to substantial improvements in energy efficiency and enabling extended operational durations. As a result, information gathering using UAVs has become increasingly cost-effective, making the coordinated operation of multiple electric UAVs practically feasible. This paper proposes a novel sliding window estimation-based high-precision perception method utilizing vision data collected from multiple UAVs. The proposed approach requires only onboard cameras and position sensors, eliminating the need for heavy and costly equipment, thereby making it suitable for multi-UAV-based surveillance. Furthermore, the algorithm is designed to robustly estimate target information even under adverse weather conditions, such as cloudy or low-visibility environments, which introduce measurement uncertainties. The accuracy and effectiveness of the proposed method are validated through both simulation studies and real-world experiments, with comparative analysis provided against existing methodologies.