Of late, man is becoming more and more removed from the machine operation and in his place, sophisticated systems for automatic control are being employed. Despite this, he has not become redundant, only his place of work has changed. Now, from a central control room he is able to observe and control many different processes by collecting information from displays, printouts and many other instruments. Concurrently, many other areas related to communications and processing have been (rapidly) developing, including visual communication by means of CRTs.
In particular, intelligent visual display units have proved to be a suitable means of communication between man and machine and have become an integral part of the control room.
Since the introduction of raster scan displays and input techniques such as: light pen, joy stick and the use of special sets of characters, the exchange of data between operator and machine has become much easier and faster.
However, computers are not the only components in an information system. Man is also an integral part of it and because of his intellectual capabilities his presence cannot be avoided. Also, optimization of results will only be possible if, besides processing human behaviour, man's interaction with machine is also taken into account.
Man is an indispensable part of many processing systems and sometimes he works as an integral part of the whole system, with information being transferred from the system to him and vice versa. Thus, special care must be taken over the way in which man-machine dialogue is performed. As human data acquisition capacity is limited, emphasis must be put on the simplification of man-machine communication.
Many studies have been performed with the objective of determining the main parameters related to human behaviour. Some of these are based on empirical measurements while others have relied on control theory. Whichever way is chosen, human behaviour is too complex to be described because of its enormous versatility. The human in a man-machine system is at the top of the hierarchy and he works as an adaptive and optimal controller. Despite these obstacles, different ways have been proposed in order to increase the knowledge of human response to external stimuli.
Many studies have been developed to relate and quantify the factors involved in man-machine dialogue. From this point of view the main factors to be considered are: task difficulties, environmental constraints and facilities, operator disposition and interest in the task being executed. Another factor of great importance is the system response time and the way that the information is presented. The main reason for this is the fact that only a limited amount of information can be held in man's short-term memory, a sort of human buffer storage. If the dialogue takes a long time, part or all of the information stored in the short-term memory may be lost due to uncoupling caused by noise, distractions or other thinking. Figure 1 shows a block diagram relating these factors.
Control theory has also been used to describe the human response to external stimuli. Different models have been suggested using open-loop, closed-loop structures or combinations of open-loop/closed-loop. These structures are classified according to their characteristics. The compensatory control is a closed-loop structure. The human controller acts in response to errors only and his aim is to minimise system error in the presence of commands and disturbances. The pursuit and precognitive controls are open-loop structures and are used when the output to be attained is known or foreseen. Normally, these structures appear together. The control is initiated and mostly accomplished through the precognitive action and then completed with compensatory operations. Figure 2 shows a compensatory control diagram.
In both models, the human operator receives external stimuli, normally optical. The operator separates the information needed for monitoring from that necessary for control purposes. Only the latter is used by the operator with the objective of control but both have influence on the operator response.
The process of taking a control action is influenced to a large extent by the quality of stimuli received. In many cases these stimuli are obtained visually from displays whose main characteristics are discussed next.
Some displays are part of a system, not only to present computer output for human consumption, but also to permit input to the computer through a light pen, a typewriter or other device.
In order to determine whether a display satisfies the above requirements, we need some criteria for evaluation of its technical features. In table 1 the main criteria for this evaluation are presented.
In addition to the above, other factors, related to the shortcomings of the interactive consoles, must be taken into account. The most relevant of these factors are listed in table 2.
Observance of the above requirements is necessary but not sufficient to accomplish an optimal man-machine dialogue. A (natural) language for this dialogue as well as reasonable software support are also required.
Use of displays in process control, contrary to use in graphical applications, is basically non-interactive, being mainly for process observation. Normally an intensive interaction between man and computer (process) occurs only in critical cases, in order to correct process failures. The languages developed to handle graphical applications, like CAD are not well tailored for process control, since they do not include any flexible way for the dynamic representation of the process variables.
The system command language is probably the most important factor in determining the effectiveness of man-machine interaction. The language must be consistent, powerful, portable, modular, easy to program and must provide facilities for the operator-machine dialogue.
From the operator's point of view, the dialogue must be accomplished by using a large number of commands with few arguments, in order to avoid overloading the operators memory. It must provide simple commands with which the operator can perform a variety of immediate checks at key points of the process being monitored, as well as the ability to access files where data or documentation related to the process are stored. The dialogue may be a question and answer dialogue with the interchange guided by the user or the system, a dialogue by menu selection or a free dialogue.
The use of menu techniques is the most appropriate for process control. It demands little time from the operator to accomplish a task, reduces the probability of error due to the restriction of the input to a small set of alternatives and it is simple for operator use with little time required for training. Its disadvantage is the limited flexibility provided, making it unsuitable for certain applications.
Free dialogue techniques are the most appropriate for accomplishing complex tasks. However, they can be used only by experienced operators, since they present a higher level of communication between operator and process.
In order to improve the dialogue, special features can be used, for example, special keyboards and the discrimination of different classes of information by using colour, different levels of brightness, upper and lower case characters and the partitioning of the screen into special areas.
In the future, the use of displays in process control will increase. Manufacturing tests, traffic control, medical monitoring and factory operations can be cited among the large number of present and future applications.
Due to the large divergence of applications, global characteristics for these systems cannot be summarised. A set of specific characteristics and values must be satisfied for each application.
With the development of semiconductor technology and the advent of microprocessors, hardware features have improved while cost has been increasing. On the other hand, software costs have been increasing, leading us to the conclusion that the use of dialogue-oriented hardware will increase considerably in the near future. This, coupled with the development of process control oriented languages and the improvement of human dynamics in man-machine interactions, will lead without doubt, to the automation of processes now regarded as unfeasible.
1. Burmeister M et al - "Grundwissen fuer den Bildschirmeinsatz in der Prozessoatenverarbeintung", Kernforschung Karlsruhe: PDV 92, September 1976.
2. Dietrich E - "New trends in the application of process computers11. Seventh Congress of the IFAC, 1978.
3. Geiser G et al - "Mensch-Machine-Kommunikation in Leitstanden", Kernforschung Karlsruhe: PDV 131, October 1977.
4. Hornung M - "Untersuchngen von Datendarstellungverfahren fuer Moderne Prozessleitwarten", Kernforschung Karlsruhe: PDV 118, January 1978.
5. Hunt E, Mokous W - "Some characteristics of human information process", Advances in information science Vol 2, Plenum Press, London.
6. Lance M, Thomaz J - "Behavioural issues in the use of interactive systems", Journal of Man-Machine Studies, 1977.
7. Mcruer D - "Human dynamics in man-machine systems" - Seventh Congress of the IFAC, 1978.
8. Rijnsdorp J, Rouse W - "Design of man-machine interfaces in process control", IFAC, 1977.
9. Vamos T - "Automatic control and artificial intelligence", Seventh Congress of the IFAC, 1978.
