From Sea to Space: Advancing Perception in Harsh Domains
ICRA 2026, June 1st, Vienna, Austria
Conference Location: Hall C1
Contact: icra-s2s-organizers@umich.edu
ICRA 2026, June 1st, Vienna, Austria
Conference Location: Hall C1
Contact: icra-s2s-organizers@umich.edu
Travel Grant Submission Deadline: March 27, 2026, 11:59PM AOE
Paper Submission Deadline: April 3rd, 2026 April 10th, 2026 11:59PM AOE*
Notification of Acceptance: April 20th, 2026 April 27th, 2026
Camera-ready Submission: May 22nd, 2026, 11:59PM AOE
Workshop schedule has been posted below!
Perception is key to enabling resilient robotic autonomy in extreme environments, underpinning critical tasks such as marine imaging, planetary exploration, and lunar in-situ resource utilization. Robotic operations in harsh environments---including marine and space domains---share many common challenges, including limited computational power, unforgiving noisy sensor measurements, sparse visual features, challenging lighting conditions, GPS-denied operations, and a lack of real-world datasets. Recent developments in foundation models, 3D neural representations, and high-fidelity simulations from the vision community hold promise for accelerating robot sensing capabilities in solving these tasks. However, unique challenges encountered underwater and in space make it challenging to adopt these advancements directly. Additionally, researchers in these different domains often operate in isolation from one another, attending different conference venues.
This workshop, titled From Sea to Space: Advancing Perception in Harsh Domains has two primary aims:
Connecting researchers working on perception in marine and space applications.
Discussing how state-of-the-art vision techniques can be adapted for perception in the field.
Through invited talks, a panel discussion, and industry spotlights, this workshop aims to inspire cross-domain collaboration and transfer of state-of-the-art developments from the vision community to robotics for harsh environments.
Schedule subject to change, all times local to Vienna.
8:30 - 8:40: Opening Remarks
Session 1
8:40 - 9:10: Tobias Fischer (Fast, Efficient, and Robust: Perception and Localisation for GPS-Denied Robotics)
9:10 - 9:15: Student Spotlight 1 Daniel Yang (Underwater 3D Reconstruction by Interleaving Multimodal SLAM and Incremental Gaussian Splatting, presented by Jungseok Hong)
9:15 - 9:45: Josh Mangelson (Towards Robust Multi-Agent Underwater Localization and Coastal Semantic Mapping )
9:45 - 9:50: Spotlight 2 Tae Ha Park (Rapid and Physically-Based Gaussian Splatting of Unknown Space Objects in Low Earth Orbit)
9:50 - 10:20: Teresa Vidal-Calleja (Spatial Perception in Marine, Orbital, and Planetary Domains)
10:20 - 11:15: Coffee Break / Poster Session 1
Session 2
11:15 - 11:45: Tat-Jun Chin (Neuromorphic Perception and Computing for Space Applications)
11:45 - 11:50: Spotlight 3 Aniket Gupta (NeuSLAM: Dense Visual SLAM on Edge Devices, presented by Hanumant Singh)
11:50 - 12:20: Annette Stahl (Resilient Perception for Field Robotics in Harsh Maritime Environments)
12:20 - 1:30: Lunch
1:30 - 1:50: Industry Spotlight
1:50 - 2:00: Early Career Spotlight Grigory Solomatov
Session 3
2:00 - 2:30: Hiro Ono (Toward Interplanetary Foundation Models: Can AI Drive a Mars Rover?)
2:30 - 2:35: Spotlight 4 Yujin Park (FootRecon: Quadrupedal Terrain Reconstruction from Sparse Foot Contacts with Geometric Prior)
2:35 - 3:05: Shehryar Khattak (Redundancy Under Scarcity: Resourceful Multi-Modal Perception from Subterranean to Celestial)
3:05 - 4:00: Coffee Break / Poster Session 2
4:00 - 4:45: Panel
4:45 - 5:00: Wrap-up & Best Poster/Best Talk Awards
Hiro Ono
National Aeronautics and Space Administration, USA
Annette Stahl
Norwegian University of Science and Technology
Shehryar Khattak
FieldAI, USA
Tobias Fischer
Queensland University of Technology, Australia
Tat-Jun Chin
University of Adelaide, Australia
Joshua Mangelson
Brigham Young University, USA
Teresa Vidal Calleja
University of Technology Sydney
Submission Portal: https://openreview.net/group?id=IEEE.org/ICRA/2026/Workshop/S2S
Submitted papers should be a maximum of 6 pages in length, not including citations/acknowledgements/appendices. Papers should be formatted using the IEEE ICRA Word/Latex templates.
The papers are non-archival.
We welcome papers on topics related to perception for field robotics, particularly focused on underwater, space, and terrestrial domains. The workshop topics include but are not limited to:
Perception for scene reconstruction in harsh environments (e.g., 3D neural scene representations, real-time 3D reconstruction, multi-sensor fusion)
Explainable and trustworthy perception systems (e.g., probabilistic methods with robust uncertainty estimation, interpretable vision models)
Perception for target tracking in low lighting (e.g., feature extraction/matching, crater identification)
Perception and learning for novel sensors (e.g., event cameras, radar, star trackers, sonar, TIR cameras)
Foundation models for perception in harsh environments (e.g., fine-tuning foundation models for marine, subterranean, or space applications)
Robust localization and mapping in degraded sensing conditions (e.g., SLAM in poor lighting/dusty environments, GPS-denied global localization, multi-sensor fusion)
Datasets and scalable simulation for closing the sim-to-real gap (e.g., marine or simulation development, synthetic datasets, systems leveraging simulated data)
Computationally-efficient perception for edge deployment (e.g., lightweight machine learning models, field reports from real-time systems deployed in harsh environments)
Enabling long-term autonomy (e.g., systems that handle sensor failure, GPS-denied global localization)
Authors of accepted papers will be expected present a poster in-person at ICRA Vienna. A select number of papers will be highlighted as "spotlight" papers, and the authors will have the opportunity to give a short minute talk about their paper during the workshop program.
If you need to contact the organizers, please email icra-s2s-organizers@umich.edu.
Katie Skinner
Professor, UMich
Giancarlo Troni
Principal Engineer, MBARI
Tat-Jun Chin
Professor, Univ. of Adelaide
Ayoung Kim
Professor, Seoul National University
Marija Popovic
Professor, TU Delft
Pushyami Kaveti
Postdoc, MBARI
Travis Driver
PhD, Georgia Tech
Levi "Veevee" Cai
PhD, MIT-WHOI Joint Program
Anja Sheppard
PhD Candidate, UMich
Onur Bagoren
PhD Candidate, UMich
Jingyu Song
PhD Candidate, UMich
Amy Phung
PhD Candidate, MIT-WHOI Joint Program
Kalin Norman
PhD Candidate, BYU