The Dynamics of Trust in a Supervisory Control Simulation

Abstract

As automated controllers supplant human intervention in controlling complex systems, the operator's role changes from that of an active controller to that of a supervisory controller, interacting with the system by engaging different degrees of automatic and manual control. Because manual and automatic control have different capabilities and limitations, improperly allocating function between automatic and manual control can have negative consequences for system performance. The operator's decision to perform a task manually or automatically depends, in part, upon the trust the operators invest in the automatic controllers (Muir, 1989). Consequently, the factors influencing trust, as well as the dynamics of trust need to be identified. This research examines changes in operators' trust during an interaction with a simulated semi-automatic pasteurization plant. The results of this research show how trust develops as the operators interact with the system, as well as how trust changes in response to faults in the system. An ARMAV time series model of trust is proposed as a first step in developing a quantitative understanding of the dynamics of trust. The ARMAV analysis supports a quantitative understanding of the “inertia” observed in the both the gradual development of trust and the decline and recovery with the occurrence of faults. The findings in this paper form a foundation on which future research might build, eventually leading to a more complete understanding of the factors mediating an operator's allocation of function in a supervisory control situation.