Multi-ASV Trajectory Planning for Enhanced Water Quality Sampling

Abstract In this paper we present, an integrated system for observation of transient freshwater phenomena, mainly harmful cyanobacteria blooms, using a fleet of Autonomous Surface Vehicles (ASVs) fitted with onboard in-situ water quality sensors. Automated water quality sampling is often done following a predetermined trajectory, aiming to achieve the most cost effective representative coverage. We build on our previous work — the skeleton-of-skeleton technique, which selects sampling points representative of the body of water. Given a shared depot, the goal of the proposed algorithm is to produce trajectories for each ASV such that each sampling point is visited only once, thereby minimizing the traversal time for each robot and optimizing the operational timeline of the entire fleet. We formulate this NP-hard problem within the framework of the Multiple Traveling Salesperson Problem (mTSP), and use heuristics to address it. Water quality data was collected through multiple field deployments. Our experiments highlight the scalability of the automated system and are foundational in developing water quality sampling strategies.

@inproceedings{SalKar+24,
  author = {Ibrahim J. Salman and Nare Karapetyan and Jason M. O'Kane
            and Ioannis Rekleitis},
  booktitle = {Proc. IEEE International Conference on Robotics and
               Biomimetics},
  title = {Multi-ASV Trajectory Planning for Enhanced Water Quality
           Sampling},
  year = {2024}
}