Conclusion - Real-Time Rendering: Computer Graphics with Control Engineering - page 119

Certain visual-based criteria may serve as guidelines determining the extent to
which individual 3D objects should be scaled. In other words, the manager will act
as a “middle man” between the control action and the rendering process, distribut-
ing the effects of the control action in an elegant way to change individual objects
within the 3D scene. This would extend the utility of the proposed control system
framework to a larger pool of applications.
The next possible extension to this research is investigating other input variables
to the plant that may be used in the control system framework. Although it may be
difficult to find many variables that are easily accessible from the rendering pipeline
and able to be changed at reasonable resolution rates, it may be possible to introduce
certain user-defined input parameters specific to certain applications. For example
instead of computer graphics rendering pipeline inputs, the control system frame-
work may include hardware-related resources such as memory and CPU utilisation
that can impact rendering process performance. Also, recent advancements in com-
puter graphics hardware and techniques may expose new input variables that may be
considered in future control system frameworks.
The experiments we conducted focused largely on verification of the key con-
cepts of introducing control principles in real-time computer graphics rendering. As
a result, the simulation environment construct was performed using available test
software. For practical applications, generic libraries should be developed so that inte-
gration into different real-time rendering software can be achieved easily. However,
this proposal can involve significant time for code verification and optimisation of
runtime efficiency.