Spatial Representation and Reasoning: Integrating Synthetic Character's Visual Perception and Symbolic Representation

摘要

We propose a framework for handling information about how objects establish relationships with each other in the space. This framework is a model to address systems composed of multiple characters endowed with qualitative spatial reasoning. From the visual perception, each agent constructs its own knowledge representation about the world. These agents use imagery to compose spaces of subsets, their spatial knowledge representation. Such spaces are used as semantic domains to modelling spatial knowledge and reasoning. This kind of representation makes it natural to use constraint programming to implement an engine for spatial reasoning.