Design and Development of a Navigation System for the Visually Impaired Using the Internet of Things

Abstract

Visually impaired individuals often face difficulties navigating public spaces with various obstacles, which can pose safety risks. This research aims to design and develop a navigation system for visually impaired persons to enhance safety and promote independence when walking on sidewalks. A compact and portable prototype device was developed in the form of a belt. The system operates by applying the YOLOv11n object detection algorithm to identify objects in the walking path, determine their positions, and provide appropriate avoidance guidance. It also integrates Internet of Things (IoT) technology, comprising a Raspberry Pi 5, a camera module, an ultrasonic sensor, and an MPU6050 sensor to detecting ground slope, along with a voice alert system. The implementation process consists of (1) Designing and developing the device, (2) Training and optimizing the YOLO model for real-world environmental conditions, and (3) Testing the performance of each device function.

Experimental results evaluating the performance of each function show that the system achieves an average object detection accuracy of 94.0%, an average avoidance guidance accuracy of 95.5%, and a slope detection accuracy of 89.0%. This research presents a potential approach for practical applications to assist visually impaired individuals.