Publication Type

Conference Paper

Version

submittedVersion

Publication Date

1-2013

Abstract

Low-vision and blind bus riders often rely on known physicallandmarks to help locate and verify bus stop locations (e.g., bysearching for a shelter, bench, newspaper bin). However, there arecurrently few, if any, methods to determine this information apriori via computational tools or services. In this paper, weintroduce and evaluate a new scalable method for collecting busstop location and landmark descriptions by combining onlinecrowdsourcing and Google Street View (GSV). We conduct andreport on three studies in particular: (i) a formative interviewstudy of 18 people with visual impairments to inform the designof our crowdsourcing tool; (ii) a comparative study examiningdifferences between physical bus stop audit data and auditsconducted virtually with GSV; and (iii) an online study of 153crowd workers on Amazon Mechanical Turk to examine thefeasibility of crowdsourcing bus stop audits using our custom toolwith GSV. Our findings reemphasize the importance of landmarksin non-visual navigation, demonstrate that GSV is a viable busstop audit dataset, and show that minimally trained crowd workerscan find and identify bus stop landmarks with 82.5% accuracyacross 150 bus stop locations (87.3% with simple quality control).

Discipline

Software Engineering

Research Areas

Software and Cyber-Physical Systems

Publication

Proceedings of ASSETS 2013

Identifier

10.1145/2717513

Publisher

ACM New York

City or Country

Bellevue, Washington, USA

Additional URL

https://doi.org/10.1145/2717513

Download

Share

COinS