A Simple Heuristically-Based Method for Expressive Stimulus-Response Animations


Stimulus-Response (S-R) animation requires that each actor recognizes the state of its environment and generates a response as a function of that state. This response becomes part of the state of the environment at the next iteration. Evidence from the psychological literature shows that certain simple motions of geometric objects can be evocative of animal-like or “animate” behaviour. Our objective is to harness this effect as an aid to animation. We present a classification of S-R systems based on 1) whether the animation is high- or low-level, 2) whether the animation is stochastic or deterministic, 3) its sensory and response capabilities, 4) whether message passing or “sensing” is used, 5) whether response is based on physics or heuristic rules, and 6) the temporal definition of the stimulus. We describe an interactive low-level deterministic system designed for the rapid prototyping of animated sequences, where the stimuli are defined over at most two previous time intervals. We model S-R relationships using “behaviour functions” which are created from simple mathematical primitives in a largely heuristic manner. Results are presented which illustrate the range of behaviours we have succeeded in synthesizing, together with comments from impartial observers.

Publication Date


Journal or Conference Title

Computer Graphics


13, Issue 3



Publisher Place

New York, NY, USA



Digital Object Identifier (DOI)


Document Type

Journal Article