Navigation App for People with Blindness and Low Vision (seas.harvard.edu)

Harvard engineers have unveiled Mobilio, an innovative smartphone app designed to assist individuals with blindness or low vision in navigating their surroundings safely and efficiently. By integrating artificial intelligence, built-in sensors, and personalized audio cues, Mobilio offers continuous guidance and obstacle avoidance, outperforming Google Maps in early tests—participants navigated 13% faster with Mobilio and reduced collisions by 41%. This significant advancement addresses the common challenges faced by users, combining reliable turn-by-turn directions and real-time path guidance into a single, accessible smartphone application. The app utilizes the smartphone's sensors, including the camera and LiDAR, alongside a custom computer-vision model trained specifically to recognize pedestrian-relevant infrastructure. This focus on the user perspective allows for more accurate navigation support than existing technologies. Notably, Mobilio employs a human-in-the-loop system that adjusts audio cues to optimize the user experience in real time. The researchers emphasize the importance of user feedback, drawn from both quantitative and qualitative experiences, in shaping the app’s development. Looking ahead, the team aims to conduct further testing in diverse environments, with aspirations for broader commercialization, potentially revolutionizing navigation aids for the visually impaired.