An IoT Machine Learning-Based Mobile Sensors Unit for Visually Impaired People

Area of Science:

• Computer Science
• Assistive Technology
• Internet of Things (IoT)

Background:

• Visually impaired individuals face significant challenges in daily navigation and interaction.
• The Internet of Things (IoT) offers potential solutions for smart city applications, including aiding those with visual impairments.
• Independent mobility is crucial for the quality of life for visually impaired people.

Purpose of the Study:

• To propose and develop a smart IoT-based mobile sensor unit for a cane to enhance independent movement for visually impaired individuals.
• To integrate obstacle detection and emergency alert functionalities into an assistive device.
• To provide a safety tracking system for visually impaired cane users and their guardians.

Main Methods:

• Development of a smart cane prototype integrating a six-axis accelerometer/gyro, ultrasonic sensors, GPS, cameras, and a microcomputer.
• Implementation of an embedded machine learning algorithm for real-time obstacle identification and user alarming.
• Creation of a companion mobile application for guardian-based user tracking via Google Maps.

Main Results:

• The smart cane successfully collects environmental and user movement data.
• The embedded machine learning algorithm effectively identifies and alerts users to detected obstacles.
• The system includes an emergency alert for falls and a tracking feature for enhanced user safety.

Conclusions:

• The proposed smart cane system demonstrates a viable solution for improving independent mobility for the visually impaired.
• Integration of IoT sensors and machine learning in assistive devices can significantly enhance user safety and autonomy.
• The developed prototype validates the system's potential for real-world application in smart city environments.