The Project R1-[ImageSpace]

Mental reasoning about spatial environments often uses spatio-analogical or quasipictorial representation structures. As human working memory for spatio-analogical knowledge processing is severely restricted in capacity, mental processes dynamically construct and explore task-sensitive representations to obtain a desired spatial information. Although models of aspects of spatial mental image processing already exist, there is no computational architecture yet that describes mental knowledge processing of spatial environments as a whole. Of special interest are the dynamic inter-operation between the mental components involved and the interaction between internal and external diagrammatic representations.

The goal of the project is to model the construction and inspection of mental representations of spatial environments and to explore these models computationally. The project aims at designing and implementing a diagrammatic processing architecture that serves as a computational description of the corresponding mental processes. We will analyze the dynamic operation of the architecture, and we will investigate the feedback characteristics between the different processing instances involved. This analysis of the model will lead to iterated design steps and extensions of the model. The resulting models will be applied in a prototypical spatial task assistance system that complements internal representations by external representations interactively to compensate mental processing restrictions.

The project will be carried out from an interdisciplinary perspective. We will first analyze empirical findings with respect to the characteristics of the knowledge involved in mental image construction. We will then design a conceptual model of human reasoning about spatial environments that is based on these empirical findings. This conceptual model will serve as the basis for the specification and design of the computational model which will be done next.

The interactions between internal representations of the environment and external representations that mimic the characteristics of mental representation structures (e.g., schematic maps) require further research. Based on the model developed by then the emphasis will be on the interaction between internal and external representations. This will be done from two perspectives. First, external representations will be integrated into the model as external extensions of the modeled internal spatioanalogical representations. Second, the model extended in this way will be investigated with respect to the prototypical application. The challenging question is, how internal images and external spatial representation can be integrated into a common framework to overcome reasoning restrictions due to capacity limitations.