When preparing to visit an unfamiliar building, visual information is often used, such as building plans, images of the building or a specific feature, or 360-degree views as in Google Street View.
Many visual cues also exist on site to enable or facilitate orientation and mobility in a building. These can also be maps, but also (door) signs, directional information using arrows or information boards.
However, this information can only be used to a limited extent or not at all by people with visual impairments. Textual descriptions of buildings and the existing building features can then be an important source of information to ensure independent and autonomous mobility for people with visual impairment. Depending on the scenario, different ways of describing spatial environments can be used: egocentric or allocentric descriptions.
With allocentric descriptions information is indicated independently of the position of a user by means of an absolute reference system. Objects are therefore described in relation to each other or in relation to the reference system.
This form of description is especially relevant for planning a visit in advance.
As a basis, a study was conducted involving people with blindness and severe visual impairment. In this study, the target group’s requirements for (allocentric) descriptions of a map in preparation for a visit to a building were surveyed. It was found that a short, structured description with the relevant information is desired by the target group.
Based on the analyzed requirements, a grammar for generating allocentric building descriptions was developed and prototypically implemented. Furthermore, a revision based on user feedback was conducted.
The resulting allocentric building descriptions are integrated into Mapable as more extensive descriptions for a building and individual floors.
In egocentric descriptions, information about objects in the indoor environment, such as stairs, toilets, doors or rooms are conveyed to the user in textual form. The position of the objects is described in relation to the position and orientation of the user. The user in this case ‚sits‘ in the center of the reference system – thereby the name ‚ego-centric‘.
This kind of description is mainly relevant on site, when traveling in a building, but could also be used for a simulation or virtual tour of the building.
For designing the egocentric textual descriptions, an analysis of similar approaches from the literature, including research prototypes and similar systems (e.g. Blindsquare), was carried. Based on this analysis, a grammar in form of rules and specifications for the textual descriptions inside of buildings was developed. The grammar is specific for the German language and was evaluated in an online study with blind users. The results of the study suggest that users could form a mental map from the provided textual descriptions. Moreover, descriptions should be consistent, having a fixed word order and should also be configurable.The egocentric textual descriptions were implemented in the prototypical app MapEar.