Feasibility Study of a "Smart" Aid for the Visually Impaired and Blind's Independent Mobility in Outdoor Environments

This paper investigates feasibilities of technological solutions for empowering visually impaired and blind people to travel in outdoor environments independently. Through related work on users'spatial models, usage of long white canes, and existing technological aid tools, an in-depth understanding of users and problems of independent travel is developed. The main constituents of the independent travel are object detection, navigation, and orientation. While long white canes are very useful for supporting visually impaired and blind people to travel independently, the usefulness is limited to near-space environments. This understanding is later validated using face-to-face interviews, online surveys, and field observations. Based on the understanding, I proposed a design to utilize machine learning mechanisms and computer vision in order to provide real-time and customized support for independent travel in conjunction of use of long white canes. With feedback from user evaluation of a lo-fiprototype of the design proposal, the design concept showed potential abilities of empowering users with travelling on their own. The proven concept initiated a start of a series of futurework for possible implementations. This paper is focused only on support for individuals with visualimpairments to navigate outdoors. Thus, indoor navigation is not specifically addressed, and deafblind people are not covered in the target audience