Computational Design of Robots
Event Type
Artificial Intelligence/Machine Learning
Deep Learning
Interest Areas
New Technologies
Research & Education
Presentation Types
In Person
Registration Categories
Full Conference Supporter
Full Conference
Exhibitor Additional Full Conference
Exhibitor Full Conference
This session WILL NOT be recorded.
TimeThursday, 11 August 20229am - 12pm PDT
LocationWest Building, Room 109-110
DescriptionHow can a computer build a robot? A bottom-up introduction to computational robot co-design. Reviews (differentiable) rigid- and soft-bodied simulation and robot control methods. Discusses heuristic search, continuous optimization, and learning-based methods for co-optimizing robots over body (form) and brain (control). Describes digital fabrication, interactive design workflows, and open problems.
Prerequisites This is a rigorous technical course aimed towards both researchers and practitioners, both in robotics and neighboring fields such as physical simulation, character animation and control, automated digital content generation, computer-aided design, learning-based methods, and computational fabrication, who are interested in learning about the computationally-driven future of robot design.
Intended Audience This course does not assume any technical knowledge about robotics, simulation, control, or any other advanced domain-specific topic. In order to allow us to cover these advanced topics, the course, does, however, expect knowledge of advanced mathematics (calculus and linear algebra), a general understanding of optimization and machine learning, and familiarity with basic physics concepts.