QUALITATIVE ROUTING INSTRUCTIONS IN INDOOR SPACE

Abstract. This paper discusses the generation of routing instructions in indoor space. We develop a user-friendly application that guides a pedestrian to reach a desired destination within a typical building floor in a simply and quickly way. Starting from a suitable navigation graph of the floor, the application considers qualitative reasoning techniques to produce indications such as “go slightly further” or “keep on the right” and adds a semantic level that provides indications on the visible landmarks along the route, such as “continue after the reception desk on the right”. This characteristic is fundamental when the users do not have a previous knowledge of the building, as they can be reassured by the presence of landmarks that help to understand if the taken direction is the correct one. The algorithm proposed on this paper is able to generate navigation instructions closer to what would be the typical indications of a human guide. We evaluate the result for some case studies of building floors.