Perception, Learning, and Navigation

   

This workshop will discuss core challenges and new solutions in perception, learning and navigation for autonomous robots operating over long terms in the wild, complex and unstructured environments. With robotics applications in the wild broadening from environmental monitoring and agriculture to autonomous driving and disaster relief, the requirement for long-term robust, adaptive, and intelligent systems becomes essential. This workshop will bring together researchers and practitioners to share new technologies in sensor processing, scene understanding, mapping, localization, and machine learning to address the unique demands of long-term operations in the wild, including varying terrain, adverse weather, dynamic lighting conditions, and seasonal changes. Our goal is to foster collaboration and move toward new frontiers in robust multimodal perception, self-supervised learning, online/continual/lifelong learning, simulation-to-reality transfer, and explainable AI to enable safe, reliable and long-term autonomous robots in the wild. This workshop will feature presentations, papers, and interactive discussion by leading experts, highlighting recent developments and future research directions for enhanced long-term performance of robotic systems in the wild.

Venue

Room Strauss 2, VIECON (formerly Messe Wien), Vienna, Austria.

Call for Papers

The workshop will cover a range of topics, including but not limited to:

  • Long-term autonomy, persistent mapping, and online/continual/lifelong learning in ever-changing environments.
  • Robust perception in harsh and dynamic conditions in the wild.
  • Sensor fusion for robust long-term state estimation and scene understanding.
  • Self-supervised/unsupervised/weakly-supervised learning in LoWi.
  • Terrain classification and traversability analysis in LoWi.
  • Sim-to-real transfer and domain adaptation in LoWi.
  • Foundational models and their applications in LoWi.
  • Safety, long-term reliability, uncertainty quantification, and interpretability in LoWi.
  • Novel sensors, long-term or multi-season datasets, and benchmarks in LoWi.
  • Reinforcement learning for navigation and interaction in dynamic, unpredictable scenarios.
  • Data generation, augmentation, and active learning for the wild.
  • Applications with a clear focus on long-term perception, learning, and navigation challenges.

Submission Guidelines

  • We invite submissions in two tracks:
    • Full papers (4 to 8 pages): Eligible for Lightning Talks, Awards, and the Journal Special Issue.
    • Extended abstracts (up to 2 pages): Accepted abstracts will be presented in the Poster Session.
  • All submissions must use the standard IEEE RAS conference template and present research relevant to the workshop’s themes.
  • Submitted papers must not be under review or have been published in another archival venue.
  • A double-blind peer-review process will be implemented. Papers will be evaluated based on their novelty, technical soundness, clarity, and relevance to the workshop.
  • Accepted papers will be hosted on the workshop website but will not appear in the official IEEE Xplore proceedings, allowing for future submission to journals or major conferences.
  • All accepted papers will be invited for either a spotlight oral or a poster presentation at the workshop.
  • A selected number of high-quality full papers will be invited to submit an extended version of their work to a Special Issue of the IEEE Transactions on Field Robotics (T-FR).
  • Papers should be submitted through the CMT portal: https://cmt3.research.microsoft.com/LoWi2026

Awards and Travel Grants

Thanks to the generous support of our sponsors, we are pleased to announce the following awards and opportunities:

  • We will present a Best Student Presentation Award and a Best Presentation Award to recognize outstanding contributions to the workshop. Winners will receive a certificate and a cash prize. The selection will be based on the quality of the paper and the presentation delivered during the workshop.
  • We are committed to fostering diversity and inclusion within the robotics community. A limited number of Travel Grants are available to support student authors, with priority given to applicants from underrepresented regions or groups.
    • Eligibility: Student authors of accepted papers.
    • Application: Details on how to apply will be sent to authors upon paper acceptance.

Important Dates

  • Paper submission deadline: March 20, 2026 Extended: April 17, 2026 11:59 p.m. (Anywhere on Earth)
  • Paper notification of acceptance: April 22, 2026 May 01, 2026
  • Workshop date: June 1, 2026

Schedule

Time Speaker Topic / Title
08:50 - 09:00 Organizers Welcome and Workshop Overview
09:00 - 09:30 Timothy Barfoot Everywhere and Everywhen: Progress on Long-Term Localization with Radar
09:30 - 10:00 Xieyuanli Chen Learning Robust and Generalizable Features for Long-term Localization
10:00 - 10:20 5 papers Lightning Talks (3 min/pers)
10:20 - 10:30 RideScan AI Startup Presentation (supported by IEEE RAS TABxStartups)
10:30 - 11:00 11 papers Coffee Break (provided by ICRA) & Poster Session A (setup)
11:00 - 11:30 Ayoung Kim Long-Term LiDAR Localization in the Wild: From Foundation Models to Ultra-Lightweight Features
11:30 - 12:00 Marija Popović Learning Robust Robot Perception in Unknown Environments
12:00 - 12:10 Raise Robotics Startup Presentation (supported by IEEE RAS TABxStartups)
12:10 - 14:00 - Lunch Break (on own) & Poster Session A (continued)
14:00 - 14:30 Teresa Vidal Calleja Multi-robot Mapping in Maritime Environments
14:30 - 15:00 Kazunori Ohno Tough Physical AI for Task Automation in Harsh Environments
15:00 - 15:20 5 papers Lightning Talks (3 min/pers)
15:20 - 15:30 ATEC (sponsor) Competition Promotion
15:30 - 16:30 12 papers Coffee Break (provided by ICRA) & Poster Session B
16:30 - 17:00 All speakers Interactive Panel Discussion
17:00 - 17:30 Organizers Best Presentation Awards and Closing Remarks

Lightning Talks and Poster Sessions Details

Note to presenters: All Lightning Talk papers are also required to present a physical poster during their assigned slot.

📂 Click to expand Session A (10:30-11:00 & 12:10-14:00)
ID Presentation Paper Title
P01 Lightning When to Map? Adaptive Switching Between Localization and SLAM in Multi-Session Systems
P02 Lightning ROS 2 Implementation of Appearance-based Visual Teach and Repeat Navigation
P03 Lightning Continual Online Backward-Compatible Learning for LiDAR Place Recognition in Adverse Weather
P04 Lightning Deep Multi-Agent Reinforcement Learning for Multi-Robot Social Navigation in Constrained Environments
P05 Lightning Edge Radar Material Classification Under Geometry Shifts
P06 Poster Only In-context Adaptation of Place Recognition through Self-supervised Learning from Video
P07 Poster Only Toward Embedded Vision-Language Perception for Long-Term Autonomous Robots via Training-Free Token Pruning
P08 Poster Only GPU-Accelerated Semantic Embedded SLAM
P09 Poster Only Multi-view 6D Pose Estimation of the Aerial Docking Device for Long-Term Drone Operation in Dynamic Environments
P10 Poster Only COMPASS: Learning Global Spatial Context for Long-Range Robot Navigation
P11 Poster Only Voxels: A Lightweight Simulation for Mobile Robotics
📂 Click to expand Session B (15:30-16:30)
ID Presentation Paper Title
P12 Lightning Overcoming Nature: Perception for Autonomous Navigation in Dense Vegetation
P13 Lightning Adaptive Gaussian Process–Based Sampling for Energy-Efficient Aquatic Sensing with Autonomous Surface Vessels
P14 Lightning Disturbance-Aware Underwater Visual-Inertial Odometry via Learned Dynamics and External Force Estimation
P15 Lightning VERTIFORMER: A Data-Efficient Multi-Task Transformer on Vertically Challenging Terrain
P16 Lightning Helhest: An Affordable and Resilient R&D Platform for Long-Term Autonomous Navigation in the Wild
P17 Poster Only Towards 3D Karst Underwater Scene Reconstruction from Rotating Sonar Data
P18 Poster Only An Open-Source LiDAR and Monocular Off-Road Autonomous Navigation Stack
P19 Poster Only Building a Robust, Autonomous Pest-Control Vehicle for Real-World Agricultural Deployment
P20 Poster Only State Corrected Predictive Preference Learning for Multimodal Robot Navigation on Uneven Terrain
P21 Poster Only Vision-Language Modeling for Natural-Language Wheel Loader Assistance in Unstructured Construction Environments
P22 Poster Only Energy-Aware NECO for Single-Pass Pixel-wise Out-of-Distribution Detection in Semantic Segmentation
P23 Poster Only Extending Operational Mission Lifetimes of Free-Flying Space Robots via Hypernetwork-Based Multi-Task GNC Controller


Speakers

Timothy Barfoot
Timothy Barfoot
University of Toronto, Canada

Xieyuanli Chen
Xieyuanli Chen
National University of Defense Technology, China

Ayoung Kim
Ayoung Kim
Seoul National University, South Korea

Marija Popović
Marija Popović
Delft University of Technology, Netherlands

Teresa Vidal Calleja
Teresa Vidal Calleja
University of Technology Sydney, Australia

Kazunori Ohno
Kazunori Ohno
Tohoku University, Japan

Title: Seven Years in the Wild: Lessons from Agricultural Robots at Scale
Update: Due to unforeseen circumstances, Jaime Pulido Fentanes will be unable to join us.
Jaime Pulido Fentanes
Jaime Pulido Fentanes
Saga Robotics, Norway

Organizers

Zhi Yan
Zhi Yan
ENSTA - Institut Polytechnique de Paris, France

François Goulette
François Goulette
ENSTA - Institut Polytechnique de Paris, France

Alexandre Chapoutot
Alexandre Chapoutot
ENSTA - Institut Polytechnique de Paris, France

David Filliat
David Filliat
AMIAD, France

Andreas Nüchter
Andreas Nüchter
University of Würzburg, Germany
Hi! PARIS International Visiting Chair, France

Tao Yang
Tao Yang
Northwestern Polytechnical University, China

Martyna Poreba
Martyna Poreba
Université Paris-Saclay, CEA, List, France

Fatemeh Rekabi Bana
Fatemeh Rekabi Bana
Durham University, UK

Jindriska Deckerova
Jindriska Deckerova
Czech Technical University in Prague, Czech Republic

Tomas Krajnik
Tomas Krajnik
Czech Technical University in Prague, Czech Republic

Sponsors

           

Contact

If you have any questions, please contact us at lowiworkshop@gmail.com.


                           


The Microsoft CMT service was used for managing the peer-reviewing process for this conference. This service was provided for free by Microsoft and they bore all expenses, including costs for Azure cloud services as well as for software development and support.