General Information
The Resilient Off-road Autonomous Robotics (ROAR) builds on the success of the ICRA 2024 Workshop on Resilient Off-road Autonomy with a refined and sharpened focus on tackling the unique challenges posed by unstructured, unpredictable, and dynamic off-road environments. These settings require advanced capabilities in perception, localization, planning, and control - areas where conventional autonomous systems often fall short.
Beyond canonical topics such as robust navigation in GPS-denied areas and terrain-adaptive control strategies, the workshop will explore cutting-edge advancements. These include multi-modal perception using novel sensor modalities (e.g., radar for robust navigation, thermal cameras for visibility in low-light conditions, and event cameras for high-speed motion detection). Advanced scene representation techniques, such as NeRF for photorealistic scene reconstruction and 3D Gaussian Splatting for efficient map representations, will also be discussed. The workshop will highlight online learning approaches to enable real-time adaptation to dynamic off-road environments and examine how foundation models can contribute to generalizable and scalable off-road autonomy. Additionally, a significant focus will be placed on practical strategies for creating comprehensive datasets and establishing robust benchmarks tailored to off-road environments, emphasizing the importance of large-scale, high-quality data for method development and training.
This will be a full-day workshop. It will feature a mix of presentations, open panel discussions, and an invited poster session. The workshop targets academics and researchers in robotics, AI, and machine learning, focusing on autonomous systems for off-road applications. It also engages industry professionals from agriculture, mining, defense, and planetary exploration, along with representatives from organizations like NASA and DARPA.
Topics Covered
- Multi-modal perception
- Self-supervised learning
- Sim-to-real transfer
- Sensor fusion techniques (modalities such as radar, thermal, and event camera).
- Learning-based methods for off-road navigation
- Online adaptation
- Traversability estimation
- Foundation models for navigation
- Heterogeneous collaborative robots
- Data collection and generation techniques for off-road environments
- Off-road driving datasets
- Cross-domain transfer for off-road applications
- Scene understanding and reconstruction (e.g. but not limited to, NeRF, Neural Representations)
- Open-set visual recognition methods under constrained conditions (image and video classification, object detection, and segmentation)
- Safe robot operation in uncertain and dynamic environments
- Planning without prior knowledge of an environment
- Approaches to off-road driving that work on multiple types of robots
- Long-term localization and mapping in complex/GPS-denied environments
- Representations of terrain that model physical properties and robot-environment interaction
Submission Instructions
- Submissions should follow the RSS 2025 paper format.
- Reviews will be conducted in a double-blind manner; please ensure your paper is properly anonymized.
- While there is no strict page limit, we recommend a length of 4–8 pages, excluding references and supplementary material.
- Papers must be submitted via OpenReview.
- Accepted papers will be presented as posters, with a select few invited for spotlight presentations.
Important Dates
- Submission Start: Apr 01 2025 12:00AM UTC-0
- Submission Deadline: May 01 2025 12:00AM UTC-0
- Acceptance Notification: May 16 2025 12:00AM UTC-0
- Camera-Ready Deadline: June 01 2025 12:00AM UTC-0
- Workshop Date: Jun 25th 2025
Speakers and Panelists
Schedule
| Time | Session |
|---|---|
| 8:00-8:15 | Poster Setup |
| 8:15-8:30 | Opening Remarks |
| 8:30-9:00 | Keynote Talk 1: Radar Localization for Driving Off the Road (Tim Barfoot) |
| 9:00-9:30 | Keynote Talk 2: Learning Robust Computer Vision in Unknown Environments (Marija Popović) |
| 9:30-10:00 | Keynote Talk 3: Future Surface and Aerial Mobility for Mars Exploration (Larry Matthies) |
| 10:00-10:30 | Coffee break and Poster Session |
| 10:30-11:00 | Keynote Talk 4: Robot Navigation in Complex Indoor and Outdoor Environment (Dinesh Manocha) |
| 11:00-11:30 | Keynote Talk 5: Robotic Autonomy in Complex Environments with Resiliency (Stuart Young) |
| 11:35-12:20 | Poster Oral Session A: Spotlight Talks. |
| 12:15-2:00 | Lunch. |
| 02:00-2:30 | Keynote Talk 6: Off Road Autonomy in Feature Sparse Environments (Anton Netchaev) |
| 2:30-3:00 | Keynote Talk 7: TBD (Byron Boots) |
| 3:00-3:30 | Poster Oral Session B: Spotlight Talks. |
| 3:30-4:00 | Coffee break and Poster Session |
| 4:00-4:30 | Keynote Talk 8: TBD (Wolfram Burgard) |
| 4:30-5:00 | Keynote Talk 9: Spatial AI and ML for Autonomous Robot Navigation in Rough Terrains (Hyun Myung) |
| 5:00-5:50 | Panel Discussion: Foundation Models and Large Language Models in Off-Road Robotics: Hype or Reality? |
| 5:50-6:00 | Closing Remarks. |
Online Participation
All sessions, including invited talks, paper presentations, and panel discussions, will be live-streamed. Recordings will be made available after the event to accommodate participants in different time zones. Moreover, online attendees will have the opportunity to actively engage in live Q&A sessions via Zoom chat.
