Massive Action Control System

Full Conference
One-Day Full Conference
Basic Conference/Exhibits Plus

Massive Action Control System concurrently controls thousands of actions of multiple characters with various motivations, feelings, and personalities.

Enhanced Life
With this system, users can easily create lifelike characters and expand player experiences. For example, non-player characters can become active in massive multi-player role-playing games.

In recent years, many interactive storytelling applications have relied on progress in animation technology to create autonomous characters. Massive Action Control System is the next step in this evolution. Its purpose is to help developers create characters that perform massive actions (such as greeting, talking a walk, or going to work), reflex actions (which require reacting to input from a user), perceiving actions (where the character perceives an object and reacts to it), active actions, and actions based on personalities or feelings.

This system can execute massive actions in multiple characters. It continuously selects appropriate fragmentary behavior-control modules, called episode trees, based on the character's inner states (motives, feelings, and personality, for example) and the state of the external world, such as nearby characters and objects.

Massive Action Control System is demonstrated via an application called Spilant World, which dislays multiple characters with multiple motivations. When a user adds a new object to the virtual world or touches an object in the virtual world, the characters recognize the object or action and autonomously start new actions. This, in turn, affects the action selection of other characters, creating an opportunity for those characters to change their actions. Users can experience the story not only by tying their actions to changes in the characters' actions, but also by allowing the effect to spread over the long term.

This system expands the possibilities for likelike narrative entertainment. Users will be able to build new narratives with lifelife interactive characters in the privacy of their own homes as well as in public spaces (airports, railway stations, shopping molls, etc.).

Katsutoki Hamana
Junichi Hoshino
Atsushi Nakano
University of Tsukuba