Knowledge Transfer Project
Computer Animation for Social Signals and Interactive Behaviors
Future intelligent environments, like simulation and training environments, require knowledge about the interaction between humans and systems. For example, police training today often employs expensive actors for simulating situations where social and cultural awareness is considered important. Communication then involves more than just verbal utterances. Non-verbal modalities like gesturing or body postures are equally important to convey the right message. Smart environments, intended for training, should therefore be able to communicate, not just by means of text or speech but also by means of non-verbal modalities. For example, by means of embodied agents that use gesturing and body language to communicate. In virtual environments like Re-lion's Virtual Infantry Trainer we expect that non-kinetic and non-verbal behavior will become an important aspect of trainings situations where it can convey signals like stress, frustration, or agreement or disagreement in dialogue. The UT has developed expertise in this field within the Gate project. We aim to transfer UT's knowledge on advanced animation, and more in particular knowledge concerning the usage and implementation techniques for the Behavior Markup Language (BML) that is currently being developed and standardized in behavior generation and behavior animation.
The plan is to deliver a concrete, operational BML realizer, operating within Re-lion's system architecture that is able to interface with other intelligent systems via the BML language, and to test and to evaluate this BML realizer in a number of test cases for police training.
Job Zwiers, Universiteit Twente