The initial prototype offers basic functionality for setting up a floor of a building and navigating through a visual interface. The desktop application dedicated to administrators allows defining a map by placing walls, windows and doors, as well as points of interest. Here, too, a navigation graph can be defined, by placing control points on the map and joining these points to each other or to points of interest through edges. The mobile application dedicated to administrators allows the definition of a correspondence between the physical and the virtual space by establishing corresponding coordinate systems in the desktop and mobile applications. This is also where the space is configured by placing the ARCore Cloud anchors. The mobile application dedicated to the user allows the selection of the desired destination point and the visualization on the 2D map (minimap) of the user's position and points of interest.
Within this activity, the space configuration components and the navigation component dedicated to the end user were evaluated. The times to define a building map, points of interest, and navigation graph, were determined. Then, the calibration of the position in the physical and the virtual space and the accuracy of the configuration of the ARCore anchors were evaluated. Regarding the navigation component, the selection time of the destination point (in the graphical interface) and the correctness of the visualization on the 2D virtual map (minimap) of the user's position and the points of interest in the vicinity were evaluated.
As part of this activity, experiments were carried out to evaluate how to configure the space (on a single floor) and free/guided navigation. The use of the free navigation functionality in a laboratory room and a large hallway and the guided navigation functionality in two scenarios (one simpler and one more complex) were tested.
This deliverable describes improvements made to the advanced prototype in terms of setting up multiple floors, adding stairs and elevators, creating a dedicated audio interface for the blind, and automatically calculating the minimum path from the user's position to the selected destination point.
As part of the technical capabilities evaluation activity of the advanced prototype, the space configuration component was tested, by measuring the map, points of interest and navigation graph definition times, in several configurations (with multiple floors). Also, the navigation component was evaluated by measuring the destination point selection time (in the tap-based audio interface, dedicated to the blind) and by checking the navigation and orientation directions according to different positions and orientations of the user relative to the graph of navigation or relative to nearby points of interest.
As part of the advanced prototype testing activity, experiments were carried out with both blindfolded and blind people. We repeated the experiments presented in Deliverable 2.3, but used the audio interface for waypoint selection and TTS guidance prompts. Also, additional experiments were conducted to evaluate the guidance functionality when using the stairs and elevator.