(Disney Meets Darwin)

Motor Control


The motor control of the collection of figures in the Character Evolution Tool uses the same scheme as the Walker. Motion is the result of multiple simultaneous sine functions. In a similar fashion to McKenna's (90) motor control for the virtual roach (yet much simpler), this system utilizes a rhythmic motor control program which consists of a combination of sinusoidal angular motions in multiple limbs. The differences from one figure's motor program to another's can vary greatly according to genetic variation. A figure's motor program is continually active at all times and cannot change during the figure's lifetime. This is a time-based motion control system, as opposed to a physical-based control system. Physical-based control systems are used in [Ngo and Marks, 93], [van de Panne and Fiume, 93], and others, in which the physical aspects of the environment determine the motions of the limbs, such as contact with the ground surface, global tilting angle of the figure, etc. My figures have no stimulus/response modeling, nor any proprioceptive senses, and so their motor programs cannot change in response to the environment. In the simplest of the articulated figure species, consisting of five limbs (figure 11), joint angle changes are determined by collections of sine functions (one sine function for each of four joints).


Figure 11 The simplest of the articulated figures. To achieve motion, the four joint angles (shown as arcs) are changed according to sine functions of varying amplitudes, frequencies, and phases.

In the more complex species which have variable morphologies, I have devised a different scheme, due to the fact that the number of limbs (and therefore joints) can vary. Also, the joint angles (being 3D) must be specified with more than one angular coordinate (such as yaw, pitch, and roll - I deal with yaw and pitch only). The scheme collapses the number of genes necessary to specify the joint angle parameters in potentially complex phenotypes, such that the species can have a fixed genome length. This scheme is described in more detail in Appendix B.


PERIODICITY


(go to beginning of document)