Qualitative spatial reasoning using orientation, distance, and path knowledge

摘要

We give an overview of an approach to qualitative spatial reasoning based on directional orientation information as available through perception processes or natural language descriptions. Qualitative orientations in 2-dimensional space are given by the relation between a point and a vector. The paper presents our basic iconic notation for spatial orientation relations that exploits the structure of the spatial domain and explores a variety of ways in which these relations can be manipulated and combined for spatial reasoning. Using this notation, we explore a method for exploiting interactions between space and movement in this space for enhancing the inferential power. Finally, the orientation-based approach is augmented by distance information, which can be mapped into position constraints and vice versa.