Virtual navigation for blind people: Transferring route knowledge to the real-World

Abstract

• Blind people can use virtual navigation to quickly learn real-world short routes. • Most users gained comprehensive knowledge of all routes within three sessions. • Virtual navigation allowed users to complete 60-meter real-world routes unassisted. • When receiving in-situ navigation assistance users rely less on prior knowledge. • Some users took advantage of the knowledge acquired virtually to recover from errors. Independent navigation is challenging for blind people, particularly in unfamiliar environments. Navigation assistive technologies try to provide additional support by guiding users or increasing their knowledge of the surroundings, but accurate solutions are still not widely available. Based on this limitation and on the fact that spatial knowledge can also be acquired indirectly (prior to navigation), we developed an interactive virtual navigation app where users can learn unfamiliar routes before physically visiting the environment. Our main research goals are to understand the acquisition of route knowledge through smartphone-based virtual navigation and how it evolves over time; its ability to support independent, unassisted real-world navigation of short routes; and its ability to improve user performance when using an accurate in-situ navigation tool (NavCog). With these goals in mind, we conducted a user study where 14 blind participants virtually learned routes at home for three consecutive days and then physically navigated them, both unassisted and with NavCog. In virtual navigation, we analyzed the evolution of route knowledge and we found that participants were able to quickly learn shorter routes and gradually increase their knowledge in both short and long routes. In the real-world, we found that users were able to take advantage of this knowledge, acquired completely through virtual navigation, to complete unassisted navigation tasks. When using NavCog, users tend to rely on the navigation system and less on their prior knowledge and therefore virtual navigation did not significantly improve users’ performance.