The Compliant Control Architecture (CCA) is a component architecture developed to enable machine learning and rich motor skills on compliant robot platforms, as part of the AMARSi project.
CCA provides a component model with an extended lifecycle to manage different machine learning modes (online learning, offline learning). Communication between components is done via data-flow. To easily manage static and dynamic connections between components, CCA provides Connectivity Skeletons to manage the connectivity and Processing Strategies to manage the control flow.
See: