A hand is used when operating most of the things around us, such as cellular phones, computer mice, cameras, packages, remote controllers, golf clubs, tennis rackets, etc. "Digital Hand" is software developed to assist design of such products. Given a 3D object model and the users' condition, it virtually reconstructs various sizes and shapes of hands and evaluates the usability of the object by simulating interaction between the hands and object during operation.
Compared with the case of a whole body (computer manikins), a digital hand has its own difficulties. Posture reconstruction accuracy is required to be at the millimeter order level, especially at fingertips. Diverse patterns of motion should be generated for a high degree-of-freedom linkage. There are more contact points with between a hand and external objects than for the rest of a whole body, so estimating a reaction force is difficult. In addition, contact, deformation, friction, and the tactile functionality to sense them need to be incorporated. Numerous contact points between hand and an object makes it difficult to estimate a reaction force.
A hand consists of basic structures such as bones, tendons, ligaments, muscles, and skin. Under the skin, soft tissues are packed, and tactile receptors can be found among them. Though these basic structures have been unraveled by medical science and anatomy, quantitative design of most of them is still unknown. In the digital hand project, we are trying to deal with various "individual differences" based on quantitative unraveling of these basic structures.
The physical interaction between a hand and an object is ultimately recognized and evaluated as "usability" or "feel of grip" in the brain.
It is difficult to simulate the cerebral nerve activities directly to estimate such "feel-of-operation". We are now tackling this highly challenging problem by developing an equivalent relationship model between the evaluations and the physical phenomenon that could be measured experimentally and simulated.