Authors: Elizabeth Gajski, Raymond Turrisi, Jungseok Hong, John Morrison, Alan Papalia, Eric Gallimore, Michael Benjamin, John Leonard

Venue: IEEE/OES Autonomous Underwater Vehicles Symposium (AUV), 2026

Status: Accepted

Abstract

Accurate navigation for teams of autonomous underwater vehicles (AUVs) in GPS-denied environments is a prerequisite for persistent multi-vehicle underwater operations. Fixed acoustic infrastructure is costly to survey and deploy; cooperative navigation offers an alternative in which vehicles serve as mobile acoustic references for one another, enabling accurate localization without pre-surveyed seafloor transponders. This paper presents a simulation and field study comparing and deploying cooperative versus non-cooperative range-aided estimators on compact, low-cost AUV hardware, with field trials in the Charles River. In simulation, three localization strategies are evaluated across single- and multi-vehicle scenarios: a range-aided Extended Kalman Filter (EKF), an incremental factor graph optimizer (GTSAM-iSAM2 with landmark ranges), and a Hybrid solver that couples iSAM2 with an EKF output layer; in two-vehicle scenarios, GTSAM-iSAM2 and Hybrid are extended to a cooperative configuration, with the EKF as the non-cooperative baseline. In simulation, GTSAM achieves lower cross-track error than the EKF across both anchor configurations. Cooperative iSAM2 on two vehicles achieves cross-track performance comparable to the non-cooperative EKF baseline, with improvements under configurations where inter-vehicle ranging compensates for poor fixed landmark geometry. Single and dual-agent cooperative field results are additionally presented, demonstrating live inter-vehicle acoustic ranging and joint factor graph construction with two simultaneously submerged vehicles. In the dual-agent field mission, online cooperative iSAM2 achieves 1.30 m terminal error after 13.8 min submerged, demonstrating that sub-2 m GPS-denied accuracy is achievable on low-cost, field-deployable AUV hardware using cooperative acoustic ranging alone.

Links

BibTeX

@inproceedings{gajski2026coop,
  title={A Simulation and Field Study Analysis of Cooperative Range-Aided
         Localization Strategies for Compact, Low-Cost AUV Teams},
  author={Gajski, Elizabeth and Turrisi, Raymond and Hong, Jungseok and Morrison, John
          and Papalia, Alan and Gallimore, Eric and Benjamin, Michael and Leonard, John},
  booktitle={IEEE/OES Autonomous Underwater Vehicles Symposium (AUV)},
  year={2026},
  organization={IEEE}
}

← Back to Publications