Geppetto - Enabling Semantic Design of Expressive Robot Behaviors

Explore an innovative system for designing expressive robot behaviors through semantic editing in this conference talk from the ACM CHI Conference on Human Factors in Computing Systems. Dive into Geppetto, a data-driven tool that combines physics-based simulation with crowd-powered semantic mapping to simplify the complex process of robot motion design. Learn how this approach enables mixed-initiative exploration of robot motions, focusing on emotionally expressive behaviors. Discover the system's effectiveness in quickly developing desirable robot behaviors compared to manual parameter editing, and understand its potential applications across various robotic platforms. Gain insights into the challenges of expressive robot design and how Geppetto's framework addresses them through semantic design and exploration techniques.

Syllabus

Geppetto Enabing Semantic Design of Expressive Robot Behaviors
Why do we need a design tool!
Taking a step beyond existing tools
Semantic mapping framework
Dataset synthesis
Relating motion parameters and expression score
Semantic editing with Geppetto
Geppetto enables creating better behaviors beyond the pre-synthesized data
Can we re-use Geppetto's framework for different robots?
Summary Semantic design and exploration of robot behaviors