
PS is excited to start a new collaboration with SMART at Saxion on the very relevant topic of AI-based robust Indoor/Outdoor Localization for autonomous mobile robots.
Ensuring stable and accurate localization when deploying robots between indoor and outdoor environments is critical for industrial applications such as transporting harvested crops, docking, navigating under solar panels, passing through tunnels, and parking in garages. These transitions pose challenges due to sudden changes in operating conditions, including variations in satellite signal quality, field of view, lighting, and movement speed. Our experience with multiple projects revealed that relying on individual ROS2 modules often results in inaccurate localization during these indoor-outdoor transitions.
Since commercial solutions for such transitions are scarce, AlFusIOn will introduce a ROS2-based framework that fuses multiple sensing and interpretation methods: LiDAR, cameras, IMU, GNSS-RTK, wheel odometry, and visual odometry, to ensure redundancy and accuracy. Integrated maps will further enhance robustness and safety by providing geometric details of transitional structures. Additionally, deep learning will classify indoor and outdoor areas, embedding this context into maps so robots can anticipate operational conditions beyond their current sensor data. This self-awareness will guide the fusion process, enabling dynamic switching between localization techniques for smooth transitions—for example, disabling GNSS-RTK during indoor movement.
PS' own Alex Chiumento will drive the UT's efforts in this exciting new adventure.
More recent news
Mon 3 Nov 2025Pervasive Systems Group Launches a Public GitHub Repository
Wed 29 Oct 2025CONGRATULATIONS TO DR. SON MINH NGUYEN
Mon 27 Oct 2025Guest Lecture by Prof. Valérie Renaudin on Positioning and Navigation for Sustainable Mobility
Thu 2 Oct 2025Remembering the Legacy of Paul Havinga in Rob Bemthuis’ PhD Journey
Wed 1 Oct 2025Jeanne Parmentier wins the Best Poster Award at iWOAR 2025