Artificial General Intelligence

18Combining AI and Sensor Fusion Technology for Improving Mobility Solutions Designed for Visually Impaired Users

M.A. Mohamed Aslam^*, G. Revathy, A. Gayathri, M. Krithika and B. M. Shruthi

Department of Computer Science and Engineering, Srinivasa Ramanujan Centre,

SASTRA Deemed University, Kumbakonam, India

Abstract

The identification of objects and distance evaluation using machine learning within visually impaired assistive devices remains a challenging process. The study evaluates assistive technologies based on machine learning and identifies solutions to address resulting obstacles. The visually impaired population faces challenges when identifying complex things outside and measuring distances in multiple lighting scenarios while operating with limited processing capabilities. Multiple analysis domains are currently studied to find solutions to these challenges. Deep learning and computer vision (CV) advancements have remarkably improved both safety and mobility abilities for people who are visually impaired. YOLO (You Only Look Once) operates in real time to find object distances while performing classification tasks. Multiple sensors working together as a single system makes distance measurements more precise across different types of lighting. Designing faster assistive technology heavily depends on improved computational performance alongside the creation of lightweight mobile-device models. The solution of problems requires computer vision specialists to work together ...