Intelligent appearing motion in virtual environments

Motion strongly influences the appearance and use of virtual environments (VE). It is not only essential for interacting with the virtual environment but also for their understanding. Motion has many facets which go beyond simply reaching a goal. Motion information, for example, often serves as the basis for inferring interrelationships between events, states and intentions of an acting entity. Observers can often identify or classify persons from afar by their movement alone using even low resolution views. In this connection it is important to identify dominant factors influencing such identification. It is also important to analyze how this understanding of motion relates to its primarily technological understanding of the structure of motion. This paper addresses some of these issues. By emphasizing the compositional aspect of motion it extends the common technological structuring approach. It does not offer a complete new solution but indicates directions for further development. As a first step, a linguistic framework is used for structuring complex motion. In this case, single movements serve as lexical elements and form a general motion vocabulary. Semantic rules define links and relationships between motion and overall goals, pragmatic rules between acting entities and motions, and syntactic rules between different single motions. By specifying the connection and links between single motions of a complex motion sequence, this compositional approach may make a description of a virtual environment more natural and inclusive. Further insights into this topic are given by Alexander & Ellis (2007).