Over the years machines of all
kinds have been improved and made more reliable. However, machines typically
operate as components of larger systems, such as transportation systems,
communication systems, manufacturing systems, defense systems, health care
systems, and so on. While many aspects of such systems can be and have been
automated, the human operator is retained in many cases. This may be because of
economics, tradition, cost, or (most likely) capabilities of the human to
perceive patterns of information and weigh subtle factors in making control
decisions which the machine cannot match.
Although the public as well as
those responsible for system operation usually demand that there be a human
operator, “human error” is a major reason for system failure. And aside from
prevention of error, getting the best performance out
of the system means that human and machine must be working together effectively — be properly “impedance matched.”
Therefore, the performance capabilities of the human
relative to those of the machine must be taken into account in system design.
Efforts to “optimize”
the human-machine interaction are meaningless in the mathematical sense of optimization,
since most important interactions between human and machine cannot be reduced
to a mathematical form, and the objective function (defining what is good) is
not easily obtained in any given context. For this reason, engineering the
human-machine interaction, much as in management or medicine, remains an art
more than a science, based on laboratory experiments and practical experience.
In the broadest sense,
engineering the human-machine interface includes all of ergonomics or human
factors engineering, and goes well beyond design of displays and control
devices. Ergonomics includes not only questions of sensory physiology, whether
or not the operator can see the displays or hear the auditory warnings, but
also questions of biomechanics
, how the body moves,
and whether or not the operator can reach and apply proper force to the
controls. It further includes the fields of operator selection and training,
human performance under stress, human factors in maintenance, and many other aspects
of the relation of the human to technology. This section focuses primarily on
human-machine interaction in control of systems.
The human-machine
interactions in control are considered in terms of Figure
6.1.1. In Figure 6.1.1a the human directly controls the machine; i.e.,
the control loop to the machine is closed through the physical sensors,
displays, human senses (visual, auditory, tactile), brain, human muscles,
control devices, and machine actuators. Figure 6.1.1b illustrates what has come
to be called a supervisory control system , wherein the human
intermittently instructs a computer as to goals, constraints, and procedures,
then turns a task over to the computer to perform automatic control for some
period of time.
Displays and control devices can
be analogic (movement signal directions and extent of control action, isomorphic
with the world, such as an automobile steering wheel or computer mouse
controls, or a moving needle or pictorial display element). Or they can be symbolic
(dedicated buttons or generalpurpose keyboard controls, icons, or alarm light
displays). In normal human discourse we use both speech (symbolic) and gestures
(analogic) and on paper we write alphanumeric text (symbolic) and draw pictures
(analogic). The system designer must decide which type of displays or controls
best suits a particular application, and/or what mix to use. The designer must
be aware of important criteria such as whether or not, for a proposed design,
changes in the displays and controls caused by the human operator correspond in
a natural and common-sense way to “more” or “less” of some variable as expected
by that operator and correspond to cultural norms (such as reading from left to
right in western countries), and whether or not the movement of the display
elements correspond geometrically to movements of the controls.
Leave a Reply