AI Based Handheld Electronic Travel Aid for Visually Impaired People

Visual Impairment is a major challenge. Blind people face significant problems in interacting with their surroundings. This paper presents an electronic travel aid for safe mobility of visually impaired people. This aid is implemented in the form of a hand-held torch. It consists of a monocular camera, an ultrasonic sensor, a raspberry pi board-based control system, and an earphone. It detects the obstacles using dual confirmation through a camera and ultrasonic sensor. The camera is used to detect and recognize the position of the object and an ultrasonic sensor is used for the depth perception. It can detect any obstacle in the path of visually impaired people. It can recognize a total of 80 indoor and outdoor objects/obstacles. It uses a simplified audio feedback to convey the details of detected obstacle/recognized object. The weight of the system is 410 grams. Total four experiments were conducted with blind-folded subjects to assess utility of this aid. The system consistently detected obstacles in the surrounding environment and helped the user to negotiate with the detected obstacle.