Workshop Engineering and Music

"Human Supervision and Control in Engineering and Music"

Johan Stahre

Human Supervision and Control in Engineering and Music

Abstract

The amateur orchestra provides an interesting version of the concept of supervisory control that also involves system development and intermittent intervention. Similarities are drawn between the orchestra and a semi-automated manufacturing system. Rapid developments in information technology into e.g. wearable or affective computers potentially means new tools for the operator as well as for the musician. In conclusion, supervisory control research should be in high demand in the near future.

The amateur orchestra and human supervisory control

Many of us have participated in the fantastic journeys of amateur or school orchestras. We have lived the hardships, the frustrations, and the conflicts between individuals and between instruments. We have felt the joy and satisfaction of gradually increasing harmony and experienced the final sweet fruit of success when the skilled conductor in the end has managed to gain full control of the orchestra.

Quite far from the delicate fine-tuning of professional musicians, the high school orchestra demonstrates not only a version of the concept of supervisory control but it also exemplifies system installation and run-in of processes to be controlled. Of course then assuming for a moment that musicians could take the role of the task-interactive computer in the supervisory control paradigm, as it was described by Sheridan [Sheridan 1992].

Orchestra development also involves the delicate process of trust-building as well as skilled disturbance handling. Trust much be high between the conductor and the musicians, as well as between musicians. As members of the orchestra are frequently replaced, competence levels easily become heterogeneous. Still, the experienced conductor has to interact with the different “agents” of his “system” through stick gestures from the stand, and also through intermittent repair work on subsystems e.g. broken strings, missing notes or technical instructions.

Great care has to be taken in establishing functional and efficient channels of communication so that the master plan, or script, can be transformed into beautiful music. For the young orchestra member, the learning process involves not only learning the piece at hand but probably also the need to acquire new skills on her instrument. Thus, the TIC autonomously refines and “reprograms” her contribution, while being supervisory controlled by the conductor. Trust is then built through positive feedback in learning and development and through the satisfaction that comes from achievement not only as an individual but as a group.

Supervisory control in the manufacturing system

Now, what if we would bring that mental model of the amateur orchestra into an industrial context. Perhaps then it would be possible to draw parallels to e.g. a semi-automated manufacturing system being installed, adjusted and fine-tuned for high-efficiency operation. Supervisory control is in manufacturing is performed by management, the system engineering team, the maintenance experts, and possibly the operators of the system (Sheridan 1992; Sanderson 1988; Stahre 1995). Basic supervisory principles apply but, as in the amateur orchestra, processes do not work right away and process sub-systems are frequently replaced or updated. That leaves management et al. with a collection of integrated processes to supervisory control, processes which also vary in functionality, trustworthiness, and efficiency.

Some processes are highly automated and controlled through information systems, basic skills not needing the maestro’s stick. Other, manual, processes need correct information to be presented to operators and maintenance personnel, indirectly instructing these “autonomous agents”. There are times when designers, management or maintenance experts themselves have to interact with the running system, doing intermittent repair work on sub-systems, in similarity with the orchestra.

Operator and supervisor tasks are getting increasingly complex as integration and automation rises. Unfortunately, the globalised economy basically give companies three choices regarding any given product in their portfolio: automate, eliminate or outsource production of the product to low-wage corners of the world. Thus, for the highly industrialized countries the average level of automation in manufacturing industries seems to be rising and the need for efficient control and operation rises with it. Information management and presentation in our selected manufacturing context is vital for system performance and new ways of working with information may be one key to successful control and operation. Personal and wearable information systems provides the means to link humans closer to information systems. Affective computers [Picard 1977] can help us in complex decisions by recognising emotional states and supporting us when we need support. But are we ready for this technology?

For the young generations (without unnecessary respect for computers or electronic devices) such tools will not be any more exotic than a microwave oven. For a generation who’s mobile telephones are marketed for their ability to intermittently control the electrical heating in the family summer house, X-by-wire technology to control your car is a natural development. Even very private areas of young people’s lives today are affected by strands of partially automated control, i.e. when young people meet over the internet and later break up their relationship with an SMS-message over their mobile phone.

Discussion and conclusion

This paper has attempted to enlighten the apparent similarities between the amateur orchestra and the shop-floor, from a human supervisory control perspective. Such comparisons can be made with numerous contexts and work places, since the increase in complexity at work-places has increased the need for humans to supervisory control many things in their surroundings. However, what does technological development mean for the manufacturing operator or the future high school orchestra member?

Is it possible that we will see “super operators” [Stahre 1998] who’s wearable information system provides them with knowledge to supervisory control a manufacturing plant single handed. Is human supervisory control possibly the only feasible solution in systems of that complexity?

Can we foresee the futuristic conductor, supervisory controlling his musicians “on-line” through devices like the MIT conductor’s jacket (Marrin 1999)? Will new affective technology enable him/her to feel the musicians pulse and emotions, suggesting a correction of the violin bow pressure to better fit the mode of the piece? Are the notes presented to the musicians via head-mounted displays and could the violins be virtual representations?

Little do we know of what possibilities that actually will be available, but a conclusion that can be drawn is that development in the manufacturing as well as in the orchestra contexts provides excellent opportunities and a greet need for advanced human supervisory control ideas.

References

Marrin, T. (1999) Inside the Conductor’s Jacket: Analysis, Interpretation, and Musical Synthesis of Expressive Gesture, MIT PhD. Thesis, Media Arts and Sciences.

Picard, R. W. (1997). Affective computing, MIT Press, Cambridge, Massachusetts, ISBN 0-262-16170-2

Sanderson, P.M. (1988) Human Supervisory Control in Discrete Manufacturing: Translating the Paradigm, In: Karwowski et al. (eds.) Analysis and Design of Hybrid Automated Systems (pp. 15-22).

Sheridan, T.B. (1992). Telerobotics, Automation, and Human Supervisory Control, MIT Press, Cambridge, Massachusetts, ISBN 0-262-19316-7.

Stahre, J. (1995). Towards Human Supervisory Control in Advanced Manufacturing Systems, Doctoral thesis, Chalmers Univ. of Tech., Göteborg.

Stahre J, and Johansson, A. (1998) Super Operator – Wearable Information System in Highly Automated Manufacturing, Preprints of: IFAC conference on man-machine systems, Kyoto, Japan.