Towards a new quality of automation in complex man-machine systems

Abstract

Supervisory control of dynamic technical systems can be changed towards a new quality of automation and cooperative man-machine decision making. This will be achieved by using technologies of knowledge-based support systems and cognitive engineering. Human-centred design approaches are needed for building such advanced automation systems for process supervision and control. Appropriate function and task allocations between human operators and computers as well as between computer control and computer support have to be found. A conceptual structure and a mathematical framework for describing the human-automation synergism in future technical systems are proposed in this paper. The conceptual structure differentiates between the general task categories of controlling and problem solving for both human and automation. It is further based on the structure of an extended User Interface Management System (UIMS) which suggests a separation into presentation, dialogue, user or operator model, application model, and tasks to be performed. Human operator behaviour, supervision and control, decision support (both application and operator oriented), and man-machine interfaces (with presentation and dialogue) are related to each other on different levels. Four examples from recent projects demonstrate the applicability of the human-centred approach towards advanced automation. Heuristic control of industrial processes as well as diagnosis of causes of technical failures in power plants are explained as examples for application oriented functionalities. Human error detection and flexible automation in aircraft guidance as well as procedural support for information search are described as possibilities for human operator oriented support.