AMP2026: A Multi-Platform Marine Robotics Dataset for Tracking and Mapping

Edwin Meriaux; Shuo Wen; David Widhalm; Zhizun Wang; Junming Shi; Mariana Sosa Guzmán; Kalvik Jakkala; Bennett Carley

cs.ROMar 4, 2026

AMP2026: A Multi-Platform Marine Robotics Dataset for Tracking and Mapping

Authors: Edwin Meriaux, Shuo Wen, David Widhalm, Zhizun Wang, Junming Shi, Mariana Sosa Guzmán, Kalvik Jakkala, Bennett Carley, +6 more

Organizations: McGill University, Montreal, Canada. · MILA - Quebec AI Institute. · Universit´e Paris-Saclay, CentraleSup´elec. · University of Minnesota. · Texas A&M University. · Woods Hole Oceanographic Institution. · Binghamton University.

Abstract

Marine environments present significant challenges for perception and autonomy due to dynamic surfaces, limited visibility, and complex interactions between aerial, surface, and submerged sensing modalities. This paper introduces the Aerial Marine Perception Dataset (AMP2026), a multi-platform marine robotics dataset collected across multiple field deployments designed to support research in two primary areas: multi-view tracking and marine environment mapping. The dataset includes synchronized data from aerial drones, boat-mounted cameras, and submerged robotic platforms, along with associated localization and telemetry information. The goal of this work is to provide a publicly available dataset enabling research in marine perception and multi-robot observation scenarios. This paper describes the data collection methodology, sensor configurations, dataset organization, and intended research tasks supported by the dataset.

Figures & tables

Panel of Figure 1 — Fig. 1 : AMP2026 dataset overview organized by task category. Colored blocks correspond to where the data of a particular class was collected, while the green symbol indicates the availability of GNSS ground truth.

Fig. 2 : Distribution of data collected in AMP2026 dataset across the different categories and classes in Figure 1

Panel of Figure 3 — Fig. 3 : Dataset collection locations for AMP2026.

Fig. 4 : Reefs off the coast of Bellairs Research Institute of McGill University in Barbados. This image was generated with inspiration from the work seen in [ 19 ]

Fig. 5 : Tracking Category 1: Top-down aerial image sample displaying a multi-drone tracking system in action to track two submerged vehicles (ID:1 and ID:2).

Fig. 6 : Tracking Category 2: Sample aerial footage of surface vessel with estimated GPS coordinates

Fig. 7 : Tracking Category 3: Sample aerial footage of submerged vehicle and swimmer with estimated GPS coordinates

Fig. 8 : Tracking Category 4: Sample underwater footage from a diver-mounted camera utilizing YOLO detection to track fellow divers and a submersible robot. The objects are being tracked with their bounding boxes shown.

Fig. 9 : Tracking Category 5: Sample underwater footage from the left camera of the underwater vehicle utilizing YOLO detection to track divers. The diver is identified as category ’person’ with confidence 0.38.

Fig. 10 : Sample composite map of Northern Bellairs reef from Figure 4 .

Fig. 11 : Sample composite map between North and South Bellairs from Figure 4 .

Explore similar work

GA3T: A Ground-Aerial Terrain Traversability Dataset for Heterogeneous Robot Teams in Unstructured Environments
May 7, 2026Siwei Cai, Knut Peterson, Quan Tran +7Multi-Robot Systems Robotic Data Collection
Drexel University, Philadelphia, USA
A2RL V\textsubscript{max}: The A2RL autonomous racing dataset for long-range, high-speed perception and multi-vehicle interaction
Jul 20, 2026Marvin Klemp, Dominic Ebner, Cornelius Schröder +15Autonomous Driving Perception Autonomous Racing
Institute of Measurement and Control Systems, Karlsruhe Institute of Technology · Code 19 Racing · Institute of Automotive Technology, Technical University of Munich +8
Seeing above the waves: A modular sensing framework for data acquisition at sea
Aug 11, 2026Jonathan E. Schmidt, Julius Wirbel, P. Nicholas Hansen +4Robotic Data Collection
Technical University of Denmark, Kongens Lyngby, Denmark

May 7, 2026J/K move · Enter open · S save

GA3T: A Ground-Aerial Terrain Traversability Dataset for Heterogeneous Robot Teams in Unstructured Environments

Siwei Cai, Knut Peterson, Quan Tran, Christian Ricks+6

Drexel University, Philadelphia, USA

Multi-Robot Systems Robotic Data Collection Robot Navigation Cooperative Perception Heterogeneous Collaborative Perception Robotic Perception

Open paper AskPDF arXiv

Heterogeneous air-ground robot teams combine complementary sensing modalities, mobility characteristics, and spatial viewpoints that can significantly enhance perception in complex outdoor environments. However, progress in multi-robot collaborative perception has been constrained by the lack of real-world datasets featuring overlapping multi-modal observations from platforms operating in unstructured terrain. We present GA3T (Ground-Aerial Team for Terrain Traversal), a real-world multi-robot collaborative perception dataset collected using a Clearpath Husky UGV and an Autel EVOII UAV across diverse unstructured environments, including forest trails, rocky paths, muddy terrain, snow piles, and grass-covered fields. The ground platform provides 3D LiDAR, stereo camera, IMU, and GPS data, while the aerial platform contributes RGB imagery, thermal/infrared observations, and GPS from a complementary overhead viewpoint, allowing for rich cross-modal and cross-view perception. The dataset is collected in 4 unique environments, with over 13,000 synchronized frames across approximately 29 minutes of operation, and includes both SAM3-based zero-shot segmentation and over 8,000 manually labeled images. A unique aspect of the dataset is its early-spring collection period, during which sparse tree canopies allow the aerial robot to partially observe the ground robot and terrain through the trees, allowing for occlusion-aware collaborative perception. Unlike prior multi-robot datasets that focus on SLAM or simulated cooperative driving, GA3T is specifically designed to support research on cross-view perception, air-ground viewpoint fusion, traversability estimation, and collaborative scene understanding in real off-road environments.

Abstract

Explore similar work

GA3T: A Ground-Aerial Terrain Traversability Dataset for Heterogeneous Robot Teams in Unstructured Environments

A2RL V\textsubscript{max}: The A2RL autonomous racing dataset for long-range, high-speed perception and multi-vehicle interaction

Seeing above the waves: A modular sensing framework for data acquisition at sea