Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Scalable Aerial GNSS Localization for Marine Robots

Shuo Wen, Edwin Meriaux, Mariana Sosa Guzmán, Charlotte Morissette, Chloe Si, Bobak Baghi, Gregory Dudek

TL;DR

The paper tackles the problem of accurately localizing near-surface marine robots when underwater GNSS is unavailable. It proposes a drone-based approach that combines data acquisition, vision-based detection via YOLO, and geometry-driven estimation to map overhead detections to GNSS coordinates, enabling both single- and multi-robot localization. The method uses data augmentation and curriculum learning to train a YOLOv11 detector, and derives a closed-form estimation chain with equations such as $D = A\tan(\alpha + \theta_y)$ and $X_o, Y_o$ mappings to surface coordinates, achieving meter-scale accuracy with real-time onboard performance on a Jetson AGX Xavier. The results demonstrate scalable, cost-effective localization for marine robotics, reducing the need for GNSS receivers on every robot while enabling cooperative localization across multiple vehicles; future work includes additional drones for verification and enhanced collaborative estimation.

Abstract

Accurate localization is crucial for water robotics, yet traditional onboard Global Navigation Satellite System (GNSS) approaches are difficult or ineffective due to signal reflection on the water's surface and its high cost of aquatic GNSS receivers. Existing approaches, such as inertial navigation, Doppler Velocity Loggers (DVL), SLAM, and acoustic-based methods, face challenges like error accumulation and high computational complexity. Therefore, a more efficient and scalable solution remains necessary. This paper proposes an alternative approach that leverages an aerial drone equipped with GNSS localization to track and localize a marine robot once it is near the surface of the water. Our results show that this novel adaptation enables accurate single and multi-robot marine robot localization.

Scalable Aerial GNSS Localization for Marine Robots

TL;DR

The paper tackles the problem of accurately localizing near-surface marine robots when underwater GNSS is unavailable. It proposes a drone-based approach that combines data acquisition, vision-based detection via YOLO, and geometry-driven estimation to map overhead detections to GNSS coordinates, enabling both single- and multi-robot localization. The method uses data augmentation and curriculum learning to train a YOLOv11 detector, and derives a closed-form estimation chain with equations such as and mappings to surface coordinates, achieving meter-scale accuracy with real-time onboard performance on a Jetson AGX Xavier. The results demonstrate scalable, cost-effective localization for marine robotics, reducing the need for GNSS receivers on every robot while enabling cooperative localization across multiple vehicles; future work includes additional drones for verification and enhanced collaborative estimation.

Abstract

Accurate localization is crucial for water robotics, yet traditional onboard Global Navigation Satellite System (GNSS) approaches are difficult or ineffective due to signal reflection on the water's surface and its high cost of aquatic GNSS receivers. Existing approaches, such as inertial navigation, Doppler Velocity Loggers (DVL), SLAM, and acoustic-based methods, face challenges like error accumulation and high computational complexity. Therefore, a more efficient and scalable solution remains necessary. This paper proposes an alternative approach that leverages an aerial drone equipped with GNSS localization to track and localize a marine robot once it is near the surface of the water. Our results show that this novel adaptation enables accurate single and multi-robot marine robot localization.
Paper Structure (18 sections, 14 equations, 10 figures)

This paper contains 18 sections, 14 equations, 10 figures.

Table of Contents

  1. Introduction
  2. Background
  3. Proposed Algorithm
  4. Data Acquisition
  5. Data Augumentation
  6. Vision-based Localization via an Aerial Observer
  7. Estimation
  8. Estimating the Position of the Object Relative to the Image Center
  9. Computing the Position
  10. Multi-Robot Localization
  11. Experiments
  12. Results
  13. Single-Robot Localization
  14. Multi-Robot Localization
  15. Sources of Error
  16. ...and 3 more sections

Figures (10)

  • Figure 1: Example of functional algorithm with a water robot and a drone.
  • Figure 2: Example of augmented images; the raw image is shown on the top left.
  • Figure 3: Sample frame showing the offset of the marine robot (Aqua2) from the center of the image.
  • Figure 4: Demonstration of the 3D geometry
  • Figure 5: Multi Robot Sample 1: Localization Haversine error over frames
  • ...and 5 more figures