Behavioral Animation - Computer Animation: Algorithms and Techniques - page 344

flock members to avoid

objects to avoid

migratory urge

flock centering

external forces
such as wind

Navigation
module

velocity matching

final desired velocity vector

physical constraints on member such
as non-penetration constraints

Pilot
module

current status of member

motion to be implemented

Flight
module

flight articulation

FIGURE 11.1
Negotiating the motion.

In addition to prioritized behaviors, the physical constraints of the flock member being modeled
need to be incorporated into the control strategy. Reynolds suggests a three-stage process consisting
of navigation, piloting, and flying (see Figure 11.1 ) . Navigation, as discussed earlier, negotiates among
competing desires and resolves them into a final desire. The pilot module incorporates this desire into
something the flock member model is capable of doing at the time, and the flight module is responsible
for the final specification of commands from the pilot module.
The navigation module arbitrates among the urges and produces a resultant directive to the member.
This information is passed to the pilot model, which instructs the flock member to react in a certain way
in order to satisfy the directive. The pilot model is responsible for incorporating the directive into the
constraints imposed on the flock member. For example, the weight, current speed and acceleration, and
internal state of the member can be taken into account at this stage. Common constraints include clamp-
ing acceleration, clamping velocity, and clamping velocity from below. The result from the pilot mod-
ule is the specific action that is to be put into effect by the flock member model. The flight module is
responsible for producing the parameters that will animate that action.

Collision avoidance
Several strategies can be used to avoid collisions. The ones mentioned here are from Reynolds's paper
[ 34 ] and from his course notes [ 35 ]. These strategies, in one way or another, model the flock member's
fov and visual processing. A trade-off must be made between the complexity of computation involved
and how effective the technique is in producing realistic and intuitive motion.