Chapter 2. Prehension

Chapter 7 - Opposition Space Phases 287
7.1 Planning an Opposition Space
In order to perform the simple task of lifting an object such as a
hammer, preparatory processes related to the organization and
planning of the upcoming movement occur in the CNS, coordinating
such a highly complex activity. In our model, this first step in the task
plan involves selecting an opposition space useful for the task. This
involves three aspects:
1) perceiving task-specific object properties,
2) selecting a grasp strategy, and
3) planning a hand location and orientation.
The choice of an opposition space depends on information
perceived about an object, including extrinsic and intrinsic properties.
The particular properties perceived are task-related. For example,
color is not particularly useful for grasping, unless we wish to select
objects only of a particular color, i.e., other than helping to distinguish
color as a feature of the object. Experimental evidence suggests that
intrinsic object properties, such as size and shape and surface spatial
density, and extrinsic object properties, such as object location and
orientation, are perceived prior to movement. Weight is estimated,
based on size/weight relationships. The particular properties are also
perceived in a hand related manner. Surfaces discounted for grasping
include inaccessible surfaces, surfaces too wide for the hand’s span,
and surfaces too narrow for the hand’s width. Part of the choosing of
an opposition space entails seeing an opposition vector which has a
hand-sized magnitude between two hand-sized surfaces that will
satisfy the required degrees of freedom of the task. Figure 7.3
summarizes this notion, showing the opposition vector and its
properties. The opposition vector has an orientation with respect to an
approach vector.
The choice of an opposition space also depends on previous
knowledge about the behavior of objects that our brains have collected
over time (phylogenetically and ontogenetically), and our ability to
predict and anticipate task-specific prehensile occurrences. For
example, the thumb and index finger do not precisely have to grasp the
object at the same time. One can anticipate how to ‘push’ the object
into the grasp. This knowledge includes anticipation of object rotations
and translations in relationship to where fingers are placed relative to
the object’s center of mass. In addition, knowledge about the cone of
friction is necessary for anticipating where to grasp and how to