Uncertainty-Aware Active Source Tracking of Marine Pollution using Unmanned Surface Vehicles

TL;DR

This work addresses autonomous localisation of marine pollution sources with unmanned surface vehicles by introducing an uncertainty-aware active tracking framework built on a high-fidelity dispersion simulator and a grid-based, categorical Bayesian tracker. The method plans informative trajectories via expected information gain and quantifies uncertainty through credible intervals, terminating when the interval width meets a threshold. Validation in a ROS-enabled, high-fidelity simulation environment shows improved accuracy and robustness over a baseline, supporting potential real-world deployment for rapid pollution response. The study lays groundwork for future hardware demonstrations and integrated exploration strategies to ensure reliable initial measurements in uncertain ocean conditions.

Abstract

This paper proposes an uncertainty-aware marine pollution source tracking framework for unmanned surface vehicles (USVs). By integrating high-fidelity marine pollution dispersion simulation with informative path planning techniques, we demonstrate effective identification of pollution sources in marine environments. The proposed approach is implemented based on Robot Operating System (ROS), processing real-time sensor data to update probabilistic source location estimates. The system progressively refines the estimation of source location while quantifying uncertainty levels in its predictions. Experiments conducted in simulated environments with varying source locations, wave conditions, and starting positions demonstrate the framework's ability to localise pollution sources with high accuracy. Results show that the proposed approach achieves reliable source localisation efficiently and outperforms the existing baseline. This work contributes to the development of full autonomous environmental monitoring capabilities essential for rapid response to marine pollution incidents.

Figures (6)

• Figure 1: Survey USV cruising in high-fidelity simulated environment, searching for pollution source.
• Figure 2: Overall workflow of the proposed framework and its associated hardware components.
• Figure 3: The dispersing pollution plume simulated by the CFD solver.
• Figure 4: Comparison of the probability models for source location estimation, in terms of error and runtime. The results are derived from trials in 4 different scenario variants.
• Figure 5: Scenario (a)-2 of the validation. The source location is at (2.5 m, 2.5 m) and the starting position is at (120 m, 120 m).