← ArchivePaper1996
Improv: A System for Scripting Interactive Actors in Virtual Worlds
Scripting system for interactive character behaviors combining procedural motion layering with behavioral state machines for real-time actors.
Abstract
Improv is a system for authoring believable interactive characters through two subsystems: an Animation Engine using procedural techniques and action layering with inverse kinematics, and a Behavior Engine with layered scripts and probabilistic decision rules. The system enables non-programmers to create real-time interactive actors that respond to users and each other while maintaining consistent personalities.
How to read this
- Category
- System: authoring interactive animated actors
- Contributions
- An Animation Engine using procedural motion, action layering, and inverse kinematics
- A Behavior Engine with layered scripts and probabilistic decision rules for personality-consistent behavior
- An authoring approach letting non-programmers script real-time actors that respond to users and each other
- Context
- Sits in the procedural-animation and interactive-character lineage and is referenced as a basis for later parameterized motion work such as Verbs and Adverbs.
- Correctness
- Presented as a working authoring system; its strength is believable real-time interaction rather than physical accuracy, and the quality of results depends on author-supplied scripts and rules.
- Clarity
- Accessible and idea-driven; a first pass conveys the two-engine architecture, a second pass clarifies the scripting and layering details.
- How to read it
- First pass for the split between motion generation and behavior scripting; second pass if you care about how layers and probabilistic rules combine to give consistent personality.
Builds on
Nothing in the archive, this is a starting point.
Related work
- Dog Code: Human to Quadruped Embodiment Using Shared Codebooks 2024 / MIG
- Character Motion Synthesis by Topology Coordinates 2009 / CGF
- Automated Extraction and Parameterization of Motions in Large Data Sets 2004 / SIGGRAPH
- Factorized Motion Diffusion for Precise and Character-Agnostic Motion Inbetweening 2024 / MIG
Keywords
This page summarises the entry and links to its original source. The archive never hosts or redistributes the publication itself.Show it in the full archive list →