This research advances robotic scene understanding for assistive technologies in built environments, enabling more independent mobility for people with physical disabilities.

• Integrates open-vocabulary semantic segmentation with uncertainty quantification
• Improves robotic navigation and interaction in complex indoor settings
• Addresses real-world variability through robust uncertainty alignment
• Enables assistive robots to recognize and respond to diverse environmental elements

Medical Impact: By enhancing smart wheelchairs and other assistive robots with better environmental perception, this technology directly supports mobility independence for people with disabilities, potentially transforming quality of life and reducing caregiver dependence.

Open-Vocabulary Semantic Segmentation with Uncertainty Alignment for Robotic Scene Understanding in Indoor Building Environments