Feature-Based Controllers for Physics-Based Character Animation
Martin De Lasa - Autodesk
April 12, 2013, 1 p.m. - April 12, 2013, 2 p.m.
MC103
Creating controllers for physics-based animation of characters is a long-standing open problem in animation and robotics. Such controllers would have numerous applications while also yielding insight into human motion. However, creating controllers remains very difficult: current approaches are either constrained to track motion capture data, are not robust, or provide limited control over style. In this talk, I'll present an approach to control of physics-based characters based on high-level features of human movement, such as center-of-mass, angular momentum, and end-effector motion. Objective terms are used to control each feature, and are combined via optimization. Using this approach, locomotion can be expressed in terms of a small number of features that control balance and end-effectors. This approach is used to build novel controllers for human balancing, standing jump, walking, and jogging. These controllers provide numerous benefits: human-like qualities such as arm-swing, heel-off, and hip-shoulder counter-rotation emerge automatically during walking; controllers are robust to changes in body parameters during movement; control parameters and goals may be modified during run-time, during control; control parameters apply to intuitive properties such as center-of-mass height; and controller may be mapped onto entirely new bipeds with different topology and mass distribution, without modifications to the controller itself. Transitions between multiple types of gaits, including walking, jumping, and jogging, emerge automatically. Controllers can traverse challenging terrain while following high-level user commands at interactive rates. Dr. Martin de Lasa is Technical Lead and Software Developer Manager at Autodesk where he leads the Media and Entertainment Division's Animation Team. Prior to joining Autodesk, he completed doctoral studies in Computer Science at the University of Toronto (2010). He also holds a M.E.Sc in Electrical Engineering from McGill University (2000), a B.E.Sc in Mechanical Engineering from the University of Western Ontario (UWO) (1998), and a B.Sc. in Computer Science (1998) from UWO. Prior to his doctoral work he was Technical Lead, Architect, and Manager, at Boston Dynamics, where he played a leading role in human simulation and robotics projects, including: Digital Biomechanics, BigDog, and LittleDog. His research interests span the areas of computer graphics and robotics, focusing on leveraging optimal control, optimization, and machine learning methods to build physically-based models of motion to ease animation/control of simulated characters and robotic systems. He is the recipient of the CAIAC doctoral dissertation award for the top Canadian dissertation in artificial intelligence in 2011.
