AI based ETA for Visually Impaired in Indoor Environment

Abstract

According to WHO-world health report on vision (2019), more than 1 billion people have vision impairment globally. The main challenge of the People with visual disability is to navigate within the crowded working place and they need to depend on others for their mobility. A novel AI (Artificial Intelligence) based Electronic Travel aid (ETA) is proposed which performs object detection and obstacle avoidance to overcome these barriers and guide the visually challenged people by giving voice commands which allows them to navigate without the requirement of any assistance.

AI-ETA consists of a Raspberry Pi-3 processor Loaded with a pre trained Convolutional Neural Network (CNN) model. The camera which is connected in the USB port of the processor captures an image in real time which is processed by the processor. CNN model identifies the obstacles with boxes and category index. The ultrasonic sensor finds the distance at which the obstacle is present. The obstacles identified and the corresponding distances are stored in the text file which is converted to voice using the Google Text To Speech module. The proposed system uses obstacle avoidance algorithm to guide visually impaired by giving proper voice commands. The system starts monitoring the heart rate of the user using heart beat sensor. If the heart rate is out of the threshold value which is set then the guardian will receive an alert message consist of heart rate detail along with their location (using GPS) information. The proposed system is tested on visually impaired people in indoor environment at different locations to prove the system effectiveness.