The context-adaptable pedestrian navigation system and usability in practical settings

Recently, the remarkable advance on mobile terminal technologies has enabled us to develop full-scaled navigation systems for pedestrians. In most systems developed so far, research has been focused on visualization effect on maps to users to guide pedestrians to their destination smoothly. The authors here present a context-adaptable pedestrian navigation system, which provides users with the preferable route, considering the conditions of each area and the objectives of the users. A graph-based approach utilizing metadata is employed to represent logical space built over the real space, and a scoring scheme is provided by partitioning target region to its edges for approximation graph model for real space, which enables to expand target region. This scheme is dependent on context to find paths to pedestrian´s destination. The usability of this system in many environments is examined by numerical simulations. The authors also discuss the possibilities of the system in practical settings in an open area, temporarily built-up sites.