Harnessing the Power of Neural Networks for Predicting Shading.

Unmanned aerial vehicles (UAVs) have emerged as indispensable tools in disaster management, providing critical support in planning, response, and recovery efforts. Integrating photovoltaic (PV) technology with UAVs offers promising opportunities to enhance their functionality and resilience in hostile environments. However, existing PV-based power management systems for UAVs face challenges related to changing light conditions and the impact of partial shading on module performance. It is shown that through embedding transistors with PV panels, the PV based power source of the UAVs can be made adaptable with the operating environment. However, existing shade detection techniques are cumbersome and inefficient.In this study, we propose a novel approach using a neural network model to accurately predict the shading percentage on PV cells, enabling dynamic power management. Through extensive experiments, we demonstrate the effectiveness of the model, achieving a high accuracy of 94% with 50 epochs and 96% with 100 epochs. This research highlights the potential of machine learning techniques in optimizing PV-based UAV power systems and provides insights for future advancements in this field. The integration of advanced power management strategies can significantly enhance the performance and adaptability of UAVs, contributing to more efficient and effective disaster response operations.