Ereigniserkennung während der Exploration linien-basierter Graphiken in virtuellen haptischen Umgebungen

Haptic representations offer access to graphical information for blind and visually impaired persons. Haptic perception requires active explorative movements, making haptic perception, in contrast to visual perception, sequential and local. In order to improve the information gain during haptic exploration and to compensate its limitations additional information can be conveyed in the auditory sensory channel, e.g., as verbal utterances. For a proactive assisting system, analyzing the user's exploratory movements can provide the basis for selecting and generating situated verbal assistance. Line-based graphics are a subset of graphical representations based on lines and objects formed from lines. They are well suited for haptic exploration in virtual environments, where the lines are explored via line following using a haptic force feedback device. Line-based graphics are a class of information graphics that can be used to depict graphics relevant to blind and visually impaired people like street maps or line charts. During haptic interaction the exploration movements are observed (by the assisting system). A knowledge based system recognizes complex exploration events from these observations. For this purpose knowledge representation and processing were developed that comprise a domain independent representation of line-based graphics and haptic exploration events. A rule based approach is used to aggregate simple, directly observed events into more complex exploration events by using knowledge about the ongoing exploration, past exploration events, the geometry of the line-based graphic and temporal reasoning. This approach offers possibilities for multimodal assistance based on the current exploration situation represented by the recognized complex exploration events. For instance the user can be pointed to parts not yet fully explored or be given overview information on the part of the graphic that the haptic exploration is focused on. Two prototypes for the domains of street maps and line graphs were developed for demonstrating the capabilities of knowledge based event recognition. Both prototypes feature components to create suitable haptic representations and domain specific knowledge from plausible data sources. Specialized 3D algorithms and spatial reasoning are applied for this purpose. Domain specific rule sets govern the recognition of domain specific exploration events. The prototype for recognizing exploration events during the haptic exploration of street maps has been successfully integrated with a component to generate verbal assistance. The resulting software was evaluated in an empirical study. The prototype for recognizing exploration events during the haptic exploration of line graphs has been extended for rule based generation of verbal assistance based on …