Internet of Things for the Visually Impaired: Design of a Wearable Obstacle-Detection and Navigation Device

Abstract

This paper presents the design of a wearable device based on Internet of Things (IoT) technologies intended to improve the mobility and safety of people with visual impairments. According to the World Health Organization, at least 2.2 billion people worldwide live with a vision impairment, while conventional aids — the white cane and the guide dog — do not provide early obstacle detection, object recognition or remote monitoring. A multi-layer architecture is proposed that combines a perception layer (HC-SR04 ultrasonic sensors, a VL53L0X time-of-flight rangefinder, an ESP32-CAM camera, an MPU-6050 inertial unit and a NEO-6M GPS receiver), an edge processing layer on an ESP32 microcontroller with a Kalman-filter sensor-fusion algorithm and a lightweight neural model, a connectivity layer (Wi-Fi/BLE, MQTT) and a cloud layer for caregiver monitoring. Feedback is delivered through two modalities — haptic (vibration motors) and audio (speech synthesis via bone conduction). Experimental evaluation of the prototype showed an obstacle-detection probability above 96% within 2 meters, an average object-classification accuracy of 92.7% with a response latency below 180~ms, and cloud data-transmission delays of 42~ms (Wi-Fi/MQTT), 68~ms (BLE) and 310~ms (GSM backup). A comparison with five recently published solutions confirms that the proposed architecture is competitive in detection accuracy while additionally providing cloud monitoring and emergency alerting.