Smartphone Application for Visually Impaired Hazard Detection on Station Platforms and its Evaluation with Congestion-Index

Abstract

In our study, we are conducting research on using of smartphones to support safe walking for visually impaired. In our previous studies, we confirmed that the system can detect steps, stairs, bicycles, and other obstacle objects which cause dangerous situations. Currently, we are extending the application to specialize in preventing the accident of falling from station platforms. The detection accuracy would be getting worser when target objects are hidden by other passengers. In other words, the crowdedness of the area affects the detection ability. One of requirements to use this app safely is to make clear the detection ability of the app with varying the crowdedness. This paper shows experimental results to evaluate the app’s performance in crowded situation using a congestion index, and the findings are discussed based on experimental results.