Sensing Environment Through Mobile: A Personalized Wearable Obstacle Detection System for Visually Impaired People

Abstract

Moving through an unknown environment becomes a real challenge for visually impaired people compared to others. The challenge of improving the mobility of visually impaired people is still exist in the world even after proposing various types of solutions. This dissertation presents a wearable obstacle detection approach which can be personalized along with the user via a mobile device. The research targets to explore the possibility of adapting sonar based echolocation concept to acquaint this approach. The time of flight (TOF) characteristic of a sonar wave has been used to identify the obstacles while several approaches are used to classify the identified obstacles. The research aims to discover the ability to identify the surrounding obstacles located underneath the waist level of the user. The direction and the distance to an obstacle are informed to the user as a vibration feedback pattern. The approach can be personalized according to characteristics of the user such as user s height, vibration sensitivity, etc. A prototype of this approach has been used to conduct user evaluation experiments which were carried out in an unknown structured environment. Further analysis on familiarization process of this system, obstacle identification and confusions, comparative analysis between white cane and the system and effects of personalization to mobility were executed in order to gauge the effectiveness and the accuracy of the approach. The results of the evaluation has been proven that the personalized system facilitate to improve the mobility of visually impaired people while giving a apparent consideration on the 3D environment.