...Fishes

``Fish'' is both singular and plural; when plural, it refers to more than one fish within the same species. The plural ``fishes'' is used when two or more species are involved [Wilson and Wilson1985].

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...method.

It supports the real-time simulation and wire-frame display (30 frames/second) of up to five swimming fish on a Silicon Graphics R4400 Indigo5#5 Extreme desktop workstation. If real-time performance is not an issue, a huge number of fish may be simulated and rendered photorealistically on such a system.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...Terzopoulos90.

An early draft of our model was developed based on a fish model that Caroline Houle (a former student at the graphics lab) built for one of her course projects. We would like to acknowledge her contribution to our work.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...agents,

An autonomous agent is an entity in a world that can act or behave on it own without explicit external control. Humans and animals are examples of natural autonomous agents.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...right''.

In fact, although often not explicitly stated, actions are most commonly defined as motor skills in robotics.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...swimming.

More specifically, referring to Table , the cross viscoelastic units maintain the streamlined integrity of the fish hence have the highest elasticity constants 47#47; The muscles have elasticity constants 49#49, and 51#51 for the remaining viscoelastic units that form other structures, such as the head.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...water.

We can simulate a fish swimming in fluid of different viscosity up to a certain degree fidelity by changing the value of 59#59. For example, 60#60 simulates fish ``swimming'' in air and 61#61 simulates fish swimming in oil-like fluid.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...diagonal.

Note that zero entries within this region will need to be stored because they can become nonzero during calculations.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...[Bathe and Wilson1976].

The symmetry and bandedness properties of the system matrix associated with the mass-spring-damper fish model would carry over to other mass-spring models [Miller1988, Terzopoulos and Waters1990, Grzeszczuk and Terzopoulos1995, Lee, Terzopoulos and Waters1995]. Hence, the skyline storage scheme is generally applicable to these models to improve computational efficiency. Note, however, that all of these prior models used simple, explicit time integration methods which do not involve matrix assembly and factorization. As we stated earlier, explicit methods are inadequate to simulate the dynamics of the fish biomechanical model because of their limited stability.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...muscles.

The swimming motor controller controls the swimming segment muscles (see Fig. ), while the turning motor controllers control the turning segment muscles.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...motion.

The wireframe and shaded fish models shown in Fig.  will be described in the next chapter.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...seen.

In our case, one can view the modeling of perceptual range as restricting the `visible' portion of the graphics database of the scene.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...internationally.

The interested reader is referred to the author's home page http://www.dgp.toronto.edu/people/tu/go-fish.html for a list of programs in which the two animations were shown.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...background).

It is easy to let the mating happen early into the animation by setting the libido parameter of the two fishes high enough.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

...intervention.

In the context of artificial life there are no constraints on what should happen when.

.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.

Xiaoyuan Tu

January 1996