Supporting knowledge-based behavior through ecological interface design

Many issues pertaining to the design of interfaces for complex human-machine systems remain to be investigated. A theoretical framework, called Ecological Interface Design (EID), was developed to address some of these topics. EID is based on Rasmussen's skills, rules, knowledge taxonomy and consists of three prescriptive design principles. The objective of these principles is to exploit the powerful capabilities of perception and action, while at the same time, providing the necessary support for more effortful and error-prone problem solving activities. This research evaluated how well an interface based on the principles of EID allows operators to cope with problem solving activities associated with unfamiliar and unanticipated events. According to EID, to properly support such knowledge-based behavior, an interface should display the physical and functional properties of the work domain in the form of a multilevel representation based on Rasmussen's abstraction hierarchy. A review of the literature in this area revealed that no experiment has ever compared a multilevel interface based on an abstraction hierarchy representation with another type of interface. An experiment was undertaken to address this important research need.The experiment was conducted within the context of DURESS (DUal REservoir System Simulation), a thermal-hydraulic process control simulation. The performance of two interfaces was compared: a traditional interface based on a physical (P) representation, and an EID interface based on a multilevel physical/functional (P + F) representation. To evaluate how well these two interfaces support knowledge-based problem solving activities, a methodology based on psychological research on the relationship between expertise and memory recall was adopted. Thus, subjects were presented with a dynamic scenario of DURESS' behavior and were asked to diagnose the event and to recall the state of the system. There were three types of events: normal, fault, and random. Two groups of subjects were used: theoretical experts in thermal-hydraulics and novices. Collectively, the findings are consistent with the following conclusion: An interface based on an abstraction hierarchy can provide more support for knowledge-based behavior than an interface based on physical variables alone because it results in a better match to the theoretical expert's mental model.U of I OnlyETDs are only available to UIUC Users without author permissio