Work

Mobility Support with Intelligent Obstacle Detection for Enhanced Safety

Year: 2024

Type: article

Abstract: In recent years, assistive technology usage among the visually impaired has risen significantly worldwide. While traditional aids like guide dogs and white canes have limitations, recent innovations like RFID-based indoor navigation systems and alternative sensory solutions show promise. Nevertheless, there is a need for a user-friendly, comprehensive system to address spatial orientation challenges for the visually impaired. This research addresses the significance of developing a deep learning-based walking assistance device for visually impaired individuals to enhance their safety during mobility. The proposed system utilizes real-time ultrasonic sensors attached to a cane to detect obstacles, thus reducing collision risks. It further offers real-time recognition and analysis of diverse obstacles, providing immediate feedback to the user. A camera distinguishes obstacle types and conveys relevant information through voice assistance. The system’s efficacy was confirmed with a 90–98% object recognition rate in tests involving various obstacles. This research holds importance in providing safe mobility, promoting independence, leveraging modern technology, and fostering social inclusion for visually impaired individuals.

Source: Optics

DOI: 10.3390/opt5040032