AI-Driven Decision Support for Spatio-Temporal Water Quality Monitoring Using Aquatic Drones

This paper presents an AI-driven framework for autonomous water quality monitoring using aquatic drones. The framework integrates dynamic drone modeling, uncertainty quantification, model predictive control (MPC) for trajectory tracking, and Gaussian process regression (GPR) for the spatiotemporal prediction of water quality parameters. The system employs adaptive sampling strategies that prioritize regions of high uncertainty to enhance data collection efficiency. Simulations conducted in ROS 2 and Gazebo demonstrate notable improvements in spatial coverage and a reduction in prediction uncertainty. This approach underscores the system’s capability to advance water quality monitoring and provides a scalable, cost-effective solution for dynamic environmental management. Hence, the method facilitates real-time decision-making and optimizes resource allocation.