Shortest Path Based Trained Indoor Smart Jacket Navigation System for Visually Impaired Person

Abstract

Visually impaired people face a lot of challenges in their day by day life. Due to blindness most of the time they depend on others for their daily movements. Many assistive technologies have been developed for blind people; most of them are expensive and designed in a complicated way. So in this paper, we represent a complete wearable navigation system for blind people based on the low expanse and truly subtle sensors, for example, Pi camera and Ultrasonic sensor. Live video analysis has been done to detect human faces and ultrasonic sensors are used to detect objects as obstacles. Raspberry Pi has been used as the main controller board. The indoor path has been pre-trained and saved in a database for blind assistance by voice command using Google Text To Speech (gTTS) API so that blind people can navigate independently. In an emergency, the blind person can seek help from the specific person by sending SOS short message service (SMS) through pressing an integrated button. This system has been tested continuously by both blindfolded and visually impaired people at various indoor locations. The outcome shows that it operates more efficiently than other assistive systems.