Introduction - Real-Time Rendering: Computer Graphics with Control Engineering - page 1

1 Introduction
1.1 BACKGROUND AND MOTIVATION
While modern rendering software claims to have controlling mechanisms that enhance
runtime performance, the mechanisms are often very primitive and inadequate. The
results of this deficiency are indeterminate drops in the visual quality of generated
imagery and frame rates that can severely affect usage experience. By applying control
theories in real-time rendering, it is possible to rectify these shortcomings altogether.
The vision is to create an intelligent rendering system that can systematically
adapt to its operating environment to produce optimum runtime performance at all
times. To our best knowledge, no commercial product exists as this work is written
and no active research is in progress in this cross-disciplinary application field.
The application of control concepts in the computer graphics software provides
new opportunities for better performance derivable from graphics hardware. Until
today, typical rendering applications struggled to utilise hardware efficiently. Much of
the burden of optimisation falls on the software programmer who must be extremely
conversant with the graphics pipeline.
The predominance of interactive computer graphics is underscored by a burgeon-
ing variety of applications in various aspects of daily life. For example, it is easy
to observe various types of interactive systems in an urban environment such as a
shopping mall or an office building. These systems range from digital signage to
projection-based displays and touch panels. At the industrial level, interactive com-
puter graphics technology powers important processes such as computer-aided design
and manufacturing, virtual prototyping, and scientific visualisation and simulation.
While customers constantly demand high quality computer-generated graphics,
the cost associated with their demands may not be within reach. To illustrate, the
price of a performance workstation is typically many times more than the cost of
a desktop PC for home use. Furthermore, mobile devices such as PDAs and cell
phones lack sufficient computing power to render high quality graphics for produc-
tivity at work.
Our research concerns a fully automated technology that circumvents the afore-
mentioned problems and allows users to enjoy high quality interactive computer
graphics on both desktop and mobile devices. The objective of this project is to
leverage earlier research on this subject and extend the work to allow a product-ready
toolkit to be developed for commercialisation opportunities.
Over the past few years, we developed a framework that realises the concept of
delivering adaptive interactive rendering through laboratory experiments, theoretical
modelling, and simulation. Our technology employs control theory and the system
identification methodology, both of which are mature fields, proven by their use in
1