Qualitative orientation reasoning in spatial multi-agent environment

Spatial agents are widely involved in robotics, simulated situations in GIS and virtual environment. Most of these agents are controlled by high-level natural language instructions or driven by limited natural language stories coming from real people. This kind of spatial information is usually qualitative rather than quantitative and is often incomplete according to the human spatial cognitive nature. This paper presents an approach to qualitative spatial orientation reasoning in 3-dimensional spatial environment. Signs in three components represent qualitative orientations in 3D space with respect to each spatial agent. By using sign algebra and efficient methods for dealing with uncertainties, the paper proposes a method for qualitative orientation reasoning between different reference agents, from which an analytical mechanism is built. The presented method can be integrated into the perception-deliberation-action control loop in spatial agent environment.