Capturing and Presenting Route Descriptions for People with Vision Loss

描述(由申请人提供)： 拟议研究的目的是为视力障碍者增加独立导航室内空间--特别是不熟悉的室内空间--的便利性。目前，如果不向视力正常的人寻求帮助，就无法做到这一点。此外，调查人员希望确定在不安装特殊建筑基础设施(例如，会说话的标志)和要求人们携带专门为此目的和仅为这一人群设计的特殊技术的情况下，可以在多大程度上做到这一点。为此目的，调查人员希望探索一种新的模式，从而建立信息数据库，以满足这一群体的具体定向和寻路需要。这些数据库将为每个可检测到的建筑特征赋予地理意义，将每个可检测到的特征视为可用于标记通往选定目的地的路线的面包屑。此外，这个建筑数据库将包含描述每个建筑空间的可达性的信息，以便可以绘制和遵循通往选定目的地的最可到达和最容易穿越的路线。为了测试这一新模式的可行性，有三个具体目标：(1)将定义可检测到的建筑特征类型，并将其与该人群的需求和多样性相匹配；(2)将利用机器人助手开发一致和准确的方法来捕获和组织这些建筑数据；以及(3)将评估使用该新系统的参与者的导航性能。为了实现第一个目标，将招募60名代表人群多样性的参与者，并要求他们在各种测试环境中描述哪些特征对他们来说是可检测和有意义的。所产生的数据库将根据与参与者技能和功能残余视力水平相关的自然发生的数据分组来组织。然后将开发基于这些数据分组的数据呈现方案，以便每个人提供的信息与其有效使用信息的能力相匹配。例如，不会向没有光感的人提供视觉提示，也不会向不阅读盲文的人提供与盲文有关的信息。第二个目标将使用机器人助理来实现，该助理将被编程为半自动建筑测量员助理。这款机器人由佐治亚理工学院医疗机器人实验室(HRL)开发，作为肢体残疾人士的助手，旨在由行动不便的人在建筑物空间中远程驾驶。它采用了光检测和测距(LIDAR)系统，该系统可以生成它所经过的空间的2D和3D渲染，并有一个摄像头来捕捉设置的图像。测量员将在台式计算机上处理和组织收集到的数据。第三个目标--评估建筑信息系统--将通过雇用30名代表人口多样性的参与者来实现。在反平衡交叉试验中，参与者使用开发的数据库的表现将与定向与机动性(O&M)指导员充分培训个人如何最好地驾驭该建筑环境的表现进行比较。假设如下：{假设1--如果为视障人群定制用于导航的输入数据，则将数据总量减少到与每个人的实际需要(即，他们的技能水平、视觉功能、触觉功能等)相关的数据量；然后，性能将接近熟悉路线的性能，在此之前，独立性能不太可能成功。}{假设2-当操作和维护专家使用开发的佐治亚理工学院HRL机器人来汇编输入到视障人群用于导航的建筑数据时，结果(收集的数据)将是可重复的、可靠的，并且在各种建筑设置中一致，如测试-重新测试试验所示。}

DESCRIPTION (provided by applicant): The purpose of the proposed research is to increase the ease of independently navigating indoor spaces-in particular unfamiliar indoor spaces-for persons with a visual impairment. Currently this cannot be done without asking a sighted person for assistance. Further, the investigators want to determine the extent to which this might be done without installing a special building infrastructure (e.g., Talking Signs) and requiring persons to carry special technologies designed just for this purpose and only for this population. To this end the investigators wish to explore a new paradigm whereby building information databases would be implemented to meet the specific orientation and wayfinding needs of this population. These databases would assign geographic meaning to every detectable building feature, treating each detectable feature as a bread crumb that can be used to mark the route to a selected destination. In addition, this building database will contain information describing the accessibility of each building space so that the most accessible and easily traversed route to a selected destination can be plotted and followed. To test the feasibility of this new paradigm, there are three specific aims: (1) detectable types of building features will be defined and matched to the needs and diversity of this population, (2) a consistent and accurate means of capturing and organizing this building data will be developed using a robotic assistant, and (3) navigation performance of participants using this new system will be evaluated. To accomplish the First aim, 60 participants representing the diversity of the population will be recruited and asked to describe, across a variety of test settings, what features are detectable and meaningful to them. The resulting database will be organized according to naturally-occurring data groupings associated with participant skills and level of functional residual vision. A data presentation schema based on these data groupings will then be developed, so that the information provided each person matches their ability to use it effectively. For instance, visual cues would not be provided to persons with no light perception, and Braille-related information would not be provided to persons who do not read Braille. The second aim will be accomplished using a robot assistant that will be programmed to work as a semi-automated building surveyor assistant. This robot, developed at the Georgia Tech Healthcare Robotics Lab (HRL) as an assistant for persons with physical disabilities, was designed to be remotely driven through building spaces by a person with limited mobility. It employs a Light Detection and Ranging (LIDAR) system that produces 2D and 3D renderings of the spaces it is driven through and has a camera to capture images of the settings. Collected data will be processed and organized on a desktop computer by the surveyor. The third aim-evaluation of the building information system-will be accomplished employing 30 participants representing the diversity of the population. In counter-balanced cross-over trials participant performance using the developed database will be compared with performance where an Orientation and Mobility (O&M) instructor has fully trained the person how best to navigate that building environment. The hypotheses are as follows: {Hypothesis 1 - If building data of import to a visually impaired population for navigtion is customized to reduce the total amount of data to just that which is relevant to each person's actual needs (i.e., their level of skill, visual function, tactile function, etc.); then performance approaching that in a familiar route will result, where independent performance was previously unlikely to be successful.} {Hypothesis 2 - When the developed Georgia Tech HRL robot is used by an O&M expert to compile building data of import to a visually impaired population for navigation, the results (data collected) will be repeatable, reliable and consistent across a variety of building settings as indicated in test-retest trials.}