A hand-centric classification of human and robot dexterous manipulation.

This work contributes to the development of a common framework for the discussion and analysis of dexterous manipulation across the human and robotic domains. An overview of previous work is first provided along with an analysis of the tradeoffs between arm and hand dexterity. A hand-centric and motion-centric manipulation classification is then presented and applied in four different ways. It is first discussed how the taxonomy can be used to identify a manipulation strategy. Then, applications for robot hand analysis and engineering design are explained. Finally, the classification is applied to three activities of daily living (ADLs) to distinguish the patterns of dexterous manipulation involved in each task. The same analysis method could be used to predict problem ADLs for various impairments or to produce a representative benchmark set of ADL tasks. Overall, the classification scheme proposed creates a descriptive framework that can be used to effectively describe hand movements during manipulation in a variety of contexts and might be combined with existing object centric or other taxonomies to provide a complete description of a specific manipulation task.