Capturing behaviour for the use of avatars in virtual environments.

Area of Science:

• Virtual Reality and Human-Computer Interaction
• Robotics and Artificial Intelligence

Background:

• Avatars, digital representations of humans in virtual environments, are typically human-controlled.
• Directly controlling every joint of a complex avatar is often impractical for users.
• Users need intuitive methods to command complex avatar behaviors via simple instructions.

Purpose of the Study:

• To outline the technological requirements for enabling avatars to interpret and execute complex behaviors from simple user commands.
• To explore methods for capturing, storing, and retrieving human behavior data for avatar use.
• To address the need for a structured approach to defining and sequencing avatar actions.

Main Methods:

• Capturing human behavior using various interfaces like haptic devices, visual observation (head movements), and natural language processing.
• Decomposing captured behaviors into atomic actions.
• Sequencing these actions using a defined grammar.
• Developing a user interface for instruction delivery to the avatar.

Main Results:

• Demonstrated feasibility of capturing and storing human behavior for avatar replication.
• Established the necessity of breaking down complex actions into simpler, grammatical components.
• Highlighted the importance of a user-friendly interface for effective avatar command.

Conclusions:

• The integration of behavior capture, grammatical sequencing, and user interface design is crucial for effective avatar control.
• Current technologies support the development of avatars that can interpret and execute complex actions based on simple instructions.
• This approach facilitates more natural and intuitive human-avatar interaction in virtual environments.