Virtual navigation for blind people: Transferring route knowledge to the real-World
Virtual navigation for blind people: Transferring route knowledge to the real-World
复制标题
DOI：
10.1016/j.ijhcs.2019.102369
复制
发表时间：
2020-03-01
影响因子：
5.4
通讯作者：
Asakawa, Chieko
中科院分区：
文献类型：
--
作者：
Guerreiro, Joao;Sato, Daisuke;Asakawa, Chieko
对于盲人来说，独立导航是一项挑战，特别是在陌生的环境中。导航辅助技术试图通过引导用户或增加他们对周围环境的了解来提供额外的支持，但准确的解决方案仍然没有广泛使用。基于这一限制以及空间知识也可以间接获得（在导航之前）的事实，我们开发了一个交互式虚拟导航应用程序，用户可以在物理访问环境之前学习不熟悉的路线。我们的主要研究目标是了解通过基于智能手机的虚拟导航获取路线知识，以及它如何随着时间的推移而演变;它支持独立，无辅助的短路线现实世界导航的能力;以及它在使用准确的原位导航工具（NavCog）时提高用户性能的能力。考虑到这些目标，我们进行了一项用户研究，其中14名盲人参与者连续三天在家中虚拟学习路线，然后在无人协助和NavCog的情况下进行物理导航。在虚拟导航中，我们分析了路线知识的演变，我们发现参与者能够快速学习较短的路线，并逐渐增加他们在短路线和长路线上的知识。在现实世界中，我们发现用户能够利用这些知识，完全通过虚拟导航获得，完成无辅助导航任务。当使用NavCog时，用户倾向于依赖导航系统，而较少依赖他们的先验知识，因此虚拟导航并没有显著提高用户的性能。
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.