A case study of experimental evaluation of adaptive interfaces

The control of complex systems involves monitoring many time varying and interacting variables. Adaptive interfaces can augment the performance of human operators by aiding in information management and other cognitive tasks in real-time. There is however no clearly established methodology for evaluating such adaptive interfaces. By comparing an adaptive interface with a non-adaptive one in the same domain one can show the relevance of any performance gains. A sample simulation environment is used for an anecdotal evaluation; a description of the underlying process for the environment is provided followed by the descriptions of the adaptive and non-adaptive interfaces. The experimental design and procedure are then detailed. In four performance conditions, the adaptive interface performed significantly better than the non-adaptive one in terms of time to achieve operating goal under fault conditions, fault detection time, number of system failures and fault detection rate, respectively.Relevance to industryAdaptive interface systems are useful in supporting the human operator in dynamic tasks which are ubiquitous in complex industrial process controls. The results of this study show that adaptive interface systems can provide improvement in system control times, fault detection and repairs, and aid the operator in vigilant tasks.