The effort of the group in this area has focused on Maxine system.
Maxine is a powerful script-directed engine for the real-time managing of virtual 3D environments and embodied agents. The system supports real-time multimodal user-virtual agent interaction through different channels: text, voice (natural language), mouse/keyboard/touch-screens and images.
Maxine 3D virtual humans are endowed with body and facial animations, lip-synch, and synthesized voice. The emotional state of the agent may vary during interaction and modulates agent’s facial and body expressions, reactions, answers and voice. Maxine virtual humans have been successfully used in different domains such as virtual real-time presenters, assistants for controlling domotic environments, pedagogical agents for teaching Computer Graphics, or virtual interpreters.