Trinh, Thanh-HaiChevaillier, PierreBarange, M.Soler, J.Loor, P. DeQuerrec, R.Sabine Coquillart and Anthony Steed and Greg Welch2013-10-312013-10-312011978-3-905674-33-01727-530Xhttps://doi.org/10.2312/EGVE/JVRC11/067-074This study is concerned with semantic modelling of virtual environments (VEs). A semantic model of a VE provides an abstract and high level representation of main aspects of the environment: ontological structures, behaviours and interactions of entities, etc. Furthermore, such a semantic model can be explored by artificial agents to exhibit human-like behaviours or to assist users in the VE. Previous research focused on formalising a knowledge layer that is a conceptual representation of scene content or application's entities. However, there still lacks of a semantic representation of spatial knowledge. This paper proposes to integrate a semantic model of directional knowledge into VEs. Such a directional model allows to specify relationships such as left , right , above or north , south that are critical in many applications of VEs (e.g., VEs for training, navigation aid systems). We focus particularly on modelling, computing, and visualising directional relationships. First, we propose a theoretical model of direction in VEs that enables the specification of direction both from a first- and third-person perspective. Second, we propose a generic architecture for modelling direction in VEs using a meta-modelling approach. Directional relationships are described in a qualitative manner and at a conceptual level, and thus are abstract from metrical details of VEs. Finally, we show how our semantic model of direction can be used in a cultural heritage application to specify behaviours of artificial agents and to visualise directional constraints.Categories and Subject Descriptors (according to ACM CCS): H.5.1 [Information Interfaces and Presentation]: Multimedia Information Systems-Artificial, augmented, and virtual realities, Evaluation/methodology