Perspectives Regarding UAS Control in Aquatic Environments (Rivers and Streams) based on Machine Learning

Abstract

In recent years, the integration of artificial intelligence (AI) and machine learning (ML) in unmanned aerial systems (UAS) has led to increased decision-making autonomy, particularly in complex and dynamic environments. This study proposes an innovative framework for the autonomous operation of UAVs in aquatic scenarios, focusing on the continuous surveillance of a moving vessel. The system uses data from multiple sensors to allow a UAV to stay within a defined perimeter around the vessel, maintain stability above the water, and automatically land on a mobile platform when necessary (e.g., in case of low battery or interference). The decision-making architecture is based on reinforcement learning algorithms for flight control and drone replacement management. The contribution of this study is to propose an intelligent and modular model for the coordination of multi-UAV systems for river missions, with direct applications in surveillance, search and rescue, and environmental monitoring.