In order to grasp an object, spatial information on the location of the object has to be transformed to adequate activations of the various muscles of the arm and hand. This transformation is similar to the one needed when pointing towards a target. The latter movements have been modeled successfully both at an abstract and a more concrete level. These models can (with the appropriate values for some parameters) generate movement trajectories that resemble experimental trajectories quite well. The aim of this project is to extend these models to the domain of grasping.
In a first attempt to model grasping we considered grasping as consisting of two pointing movements: one of the thumb and one of the finger. Qualitatively, this approach was found to be useful to explain many characteristics of grasping. To be able to make quantitative predictions, however, the constraints of the task have to be formulated in terms of the parameters of the pointing models.
We will examine the influence of variations of the constraints on both the actual and modeled movement trajectories. Two categories of constraints will be studied. The first category consists of constraints on the movement of an individual digit. Examples of such constraints are the orientation and the size of the surface that is to be grasped. The second category of constraints is caused by the fact that the finger and thumb cannot move independently. An example of such a constraint is the maximal distance between finger and thumb. |
