Toward human-machine cooperation

In human-machine systems, human activities are mainly oriented toward decision-making, encompassing monitoring and fault detection, fault anticipation, diagnosis and prognosis, and fault prevention and recovery. The objectives of this decision-making are related to human-machine system performance (production quantity and quality) as well as to overall system safety. In this context, human operators often play a double role: one negative in that they may perform unsafe or erroneous actions affecting the process, and one positive in that they are able to detect, prevent or recover an unsafe process behavior caused by another operator or by automated decisionmakers. This pluridisciplinary study combines two approaches to humanmachine systems: a human engineering approach, aimed at designing dedicated assistance tools for human operators and integrating them into human activities through human-machine cooperation, and an approach centered on cognitive psychology and ergonomics, which attempts to analyze the human activities in terms of the need for and use of such tools. This paper first focuses on parameters related to human-machine interaction, which have an influence on safety (e.g., degree of automation, system complexity, human complexity when dealing with normative and erroneous behaviors). The concept of cooperation is then introduced in response to safety concerns, and a framework for implementing cooperation is proposed. Examples in Air Traffic Control and in Telecommunication networks are used to illustrate our proposal.